* [net-next 01/14] net/mlx5e: Remove non-essential TLS SQ state bit
2021-04-16 18:54 [pull request][net-next 00/14] mlx5 updates 2021-04-16 Saeed Mahameed
@ 2021-04-16 18:54 ` Saeed Mahameed
2021-04-17 0:00 ` patchwork-bot+netdevbpf
2021-04-16 18:54 ` [net-next 02/14] net/mlx5e: Cleanup unused function parameter Saeed Mahameed
` (12 subsequent siblings)
13 siblings, 1 reply; 16+ messages in thread
From: Saeed Mahameed @ 2021-04-16 18:54 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski; +Cc: Tariq Toukan, netdev, Saeed Mahameed
From: Tariq Toukan <tariqt@nvidia.com>
Maintaining an SQ state bit to indicate TLS support
has no real need, a simple and fast test [1] for the SKB is
almost equally good.
[1] !skb->sk || !tls_is_sk_tx_device_offloaded(skb->sk)
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en.h | 1 -
.../net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h | 8 +++-----
drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 2 --
3 files changed, 3 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index e1c51eabe8fe..cb4e7aaa4f8a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -325,7 +325,6 @@ enum {
MLX5E_SQ_STATE_RECOVERING,
MLX5E_SQ_STATE_IPSEC,
MLX5E_SQ_STATE_AM,
- MLX5E_SQ_STATE_TLS,
MLX5E_SQ_STATE_VLAN_NEED_L2_INLINE,
MLX5E_SQ_STATE_PENDING_XSK_TX,
};
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
index cc0efac7b812..cc2851ecd512 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
@@ -123,11 +123,9 @@ static inline bool mlx5e_accel_tx_begin(struct net_device *dev,
mlx5e_udp_gso_handle_tx_skb(skb);
#ifdef CONFIG_MLX5_EN_TLS
- if (test_bit(MLX5E_SQ_STATE_TLS, &sq->state)) {
- /* May send SKBs and WQEs. */
- if (unlikely(!mlx5e_tls_handle_tx_skb(dev, sq, skb, &state->tls)))
- return false;
- }
+ /* May send SKBs and WQEs. */
+ if (unlikely(!mlx5e_tls_handle_tx_skb(dev, sq, skb, &state->tls)))
+ return false;
#endif
#ifdef CONFIG_MLX5_EN_IPSEC
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 6847e7b909a5..64d6c0fd92bf 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -1134,8 +1134,6 @@ static int mlx5e_alloc_txqsq(struct mlx5e_channel *c,
set_bit(MLX5E_SQ_STATE_VLAN_NEED_L2_INLINE, &sq->state);
if (MLX5_IPSEC_DEV(c->priv->mdev))
set_bit(MLX5E_SQ_STATE_IPSEC, &sq->state);
- if (mlx5_accel_is_tls_device(c->priv->mdev))
- set_bit(MLX5E_SQ_STATE_TLS, &sq->state);
if (param->is_mpw)
set_bit(MLX5E_SQ_STATE_MPWQE, &sq->state);
sq->stop_room = param->stop_room;
--
2.30.2
^ permalink raw reply related [flat|nested] 16+ messages in thread* Re: [net-next 01/14] net/mlx5e: Remove non-essential TLS SQ state bit
2021-04-16 18:54 ` [net-next 01/14] net/mlx5e: Remove non-essential TLS SQ state bit Saeed Mahameed
@ 2021-04-17 0:00 ` patchwork-bot+netdevbpf
0 siblings, 0 replies; 16+ messages in thread
From: patchwork-bot+netdevbpf @ 2021-04-17 0:00 UTC (permalink / raw)
To: Saeed Mahameed; +Cc: davem, kuba, tariqt, netdev, saeedm
Hello:
This series was applied to netdev/net-next.git (refs/heads/master):
On Fri, 16 Apr 2021 11:54:17 -0700 you wrote:
> From: Tariq Toukan <tariqt@nvidia.com>
>
> Maintaining an SQ state bit to indicate TLS support
> has no real need, a simple and fast test [1] for the SKB is
> almost equally good.
>
> [1] !skb->sk || !tls_is_sk_tx_device_offloaded(skb->sk)
>
> [...]
Here is the summary with links:
- [net-next,01/14] net/mlx5e: Remove non-essential TLS SQ state bit
https://git.kernel.org/netdev/net-next/c/2f014f4016db
- [net-next,02/14] net/mlx5e: Cleanup unused function parameter
https://git.kernel.org/netdev/net-next/c/8668587a33b9
- [net-next,03/14] net/mlx5e: TX, Inline TLS skb check
https://git.kernel.org/netdev/net-next/c/b6b3ad2175c8
- [net-next,04/14] net/mlx5e: TX, Inline function mlx5e_tls_handle_tx_wqe()
https://git.kernel.org/netdev/net-next/c/72f6f2f8d6aa
- [net-next,05/14] net/mlx5e: kTLS, Add resiliency to RX resync failures
https://git.kernel.org/netdev/net-next/c/e9ce991bce5b
- [net-next,06/14] net/mlx5e: Allow mlx5e_safe_switch_channels to work with channels closed
https://git.kernel.org/netdev/net-next/c/6cad120d9e62
- [net-next,07/14] net/mlx5e: Use mlx5e_safe_switch_channels when channels are closed
https://git.kernel.org/netdev/net-next/c/69cc4185dcba
- [net-next,08/14] net/mlx5e: Refactor on-the-fly configuration changes
https://git.kernel.org/netdev/net-next/c/b3b886cf965d
- [net-next,09/14] net/mlx5e: Cleanup safe switch channels API by passing params
https://git.kernel.org/netdev/net-next/c/94872d4ef9c0
- [net-next,10/14] net/mlx5: Allocate FC bulk structs with kvzalloc() instead of kzalloc()
https://git.kernel.org/netdev/net-next/c/5cec6de0ae09
- [net-next,11/14] net/mlx5: Add register layout to support extended link state
https://git.kernel.org/netdev/net-next/c/36830159acbe
- [net-next,12/14] net/mlx5e: Add ethtool extended link state
https://git.kernel.org/netdev/net-next/c/b3446acb2b9a
- [net-next,13/14] net/mlx5: Add helper to initialize 1PPS
https://git.kernel.org/netdev/net-next/c/302522e67c70
- [net-next,14/14] net/mlx5: Enhance diagnostics info for TX/RX reporters
https://git.kernel.org/netdev/net-next/c/95742c1cc59d
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
^ permalink raw reply [flat|nested] 16+ messages in thread
* [net-next 02/14] net/mlx5e: Cleanup unused function parameter
2021-04-16 18:54 [pull request][net-next 00/14] mlx5 updates 2021-04-16 Saeed Mahameed
2021-04-16 18:54 ` [net-next 01/14] net/mlx5e: Remove non-essential TLS SQ state bit Saeed Mahameed
@ 2021-04-16 18:54 ` Saeed Mahameed
2021-04-16 18:54 ` [net-next 03/14] net/mlx5e: TX, Inline TLS skb check Saeed Mahameed
` (11 subsequent siblings)
13 siblings, 0 replies; 16+ messages in thread
From: Saeed Mahameed @ 2021-04-16 18:54 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski; +Cc: Tariq Toukan, netdev, Saeed Mahameed
From: Tariq Toukan <tariqt@nvidia.com>
Socket parameter is not used in accel_rule_init(), remove it.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c
index 8c0f78c09215..76fd4b230003 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c
@@ -119,8 +119,7 @@ static void accel_rule_handle_work(struct work_struct *work)
complete(&priv_rx->add_ctx);
}
-static void accel_rule_init(struct accel_rule *rule, struct mlx5e_priv *priv,
- struct sock *sk)
+static void accel_rule_init(struct accel_rule *rule, struct mlx5e_priv *priv)
{
INIT_WORK(&rule->work, accel_rule_handle_work);
rule->priv = priv;
@@ -618,7 +617,7 @@ int mlx5e_ktls_add_rx(struct net_device *netdev, struct sock *sk,
init_completion(&priv_rx->add_ctx);
- accel_rule_init(&priv_rx->rule, priv, sk);
+ accel_rule_init(&priv_rx->rule, priv);
resync = &priv_rx->resync;
resync_init(resync, priv);
tls_offload_ctx_rx(tls_ctx)->resync_async = &resync->core;
--
2.30.2
^ permalink raw reply related [flat|nested] 16+ messages in thread* [net-next 03/14] net/mlx5e: TX, Inline TLS skb check
2021-04-16 18:54 [pull request][net-next 00/14] mlx5 updates 2021-04-16 Saeed Mahameed
2021-04-16 18:54 ` [net-next 01/14] net/mlx5e: Remove non-essential TLS SQ state bit Saeed Mahameed
2021-04-16 18:54 ` [net-next 02/14] net/mlx5e: Cleanup unused function parameter Saeed Mahameed
@ 2021-04-16 18:54 ` Saeed Mahameed
2021-04-16 18:54 ` [net-next 04/14] net/mlx5e: TX, Inline function mlx5e_tls_handle_tx_wqe() Saeed Mahameed
` (10 subsequent siblings)
13 siblings, 0 replies; 16+ messages in thread
From: Saeed Mahameed @ 2021-04-16 18:54 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski; +Cc: Tariq Toukan, netdev, Saeed Mahameed
From: Tariq Toukan <tariqt@nvidia.com>
When TLS is supported and enabled, every transmitted packet is tested
to identify if TLS offload is required.
Take the early-return condition into an inline function, to save
the overhead of a function call for non-TLS packets.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h | 5 +++--
drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c | 3 ---
drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h | 6 ++++++
3 files changed, 9 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
index cc2851ecd512..043c86c52798 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
@@ -124,8 +124,9 @@ static inline bool mlx5e_accel_tx_begin(struct net_device *dev,
#ifdef CONFIG_MLX5_EN_TLS
/* May send SKBs and WQEs. */
- if (unlikely(!mlx5e_tls_handle_tx_skb(dev, sq, skb, &state->tls)))
- return false;
+ if (mlx5e_tls_skb_offloaded(skb))
+ if (unlikely(!mlx5e_tls_handle_tx_skb(dev, sq, skb, &state->tls)))
+ return false;
#endif
#ifdef CONFIG_MLX5_EN_IPSEC
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c
index 2b51d3222ca1..97cbea7ed048 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c
@@ -263,9 +263,6 @@ bool mlx5e_tls_handle_tx_skb(struct net_device *netdev, struct mlx5e_txqsq *sq,
int datalen;
u32 skb_seq;
- if (!skb->sk || !tls_is_sk_tx_device_offloaded(skb->sk))
- return true;
-
datalen = skb->len - (skb_transport_offset(skb) + tcp_hdrlen(skb));
if (!datalen)
return true;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h
index 9923132c9440..5c3443200fd6 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h
@@ -47,6 +47,12 @@ u16 mlx5e_tls_get_stop_room(struct mlx5_core_dev *mdev, struct mlx5e_params *par
bool mlx5e_tls_handle_tx_skb(struct net_device *netdev, struct mlx5e_txqsq *sq,
struct sk_buff *skb, struct mlx5e_accel_tx_tls_state *state);
+
+static inline bool mlx5e_tls_skb_offloaded(struct sk_buff *skb)
+{
+ return skb->sk && tls_is_sk_tx_device_offloaded(skb->sk);
+}
+
void mlx5e_tls_handle_tx_wqe(struct mlx5e_txqsq *sq, struct mlx5_wqe_ctrl_seg *cseg,
struct mlx5e_accel_tx_tls_state *state);
--
2.30.2
^ permalink raw reply related [flat|nested] 16+ messages in thread* [net-next 04/14] net/mlx5e: TX, Inline function mlx5e_tls_handle_tx_wqe()
2021-04-16 18:54 [pull request][net-next 00/14] mlx5 updates 2021-04-16 Saeed Mahameed
` (2 preceding siblings ...)
2021-04-16 18:54 ` [net-next 03/14] net/mlx5e: TX, Inline TLS skb check Saeed Mahameed
@ 2021-04-16 18:54 ` Saeed Mahameed
2021-04-16 18:54 ` [net-next 05/14] net/mlx5e: kTLS, Add resiliency to RX resync failures Saeed Mahameed
` (9 subsequent siblings)
13 siblings, 0 replies; 16+ messages in thread
From: Saeed Mahameed @ 2021-04-16 18:54 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski; +Cc: Tariq Toukan, netdev, Saeed Mahameed
From: Tariq Toukan <tariqt@nvidia.com>
When TLS is supported, WQE ctrl segment of every transmitted packet
is updated with the (possibly empty, for non-TLS packets) TISN field.
Take this one-liner function into the header file and inline it,
to save the overhead of a function call per packet.
While here, remove unused function parameter.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h | 2 +-
.../net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c | 6 ------
.../net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h | 8 ++++++--
3 files changed, 7 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
index 043c86c52798..00af0b831a28 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
@@ -185,7 +185,7 @@ static inline void mlx5e_accel_tx_finish(struct mlx5e_txqsq *sq,
struct mlx5_wqe_inline_seg *inlseg)
{
#ifdef CONFIG_MLX5_EN_TLS
- mlx5e_tls_handle_tx_wqe(sq, &wqe->ctrl, &state->tls);
+ mlx5e_tls_handle_tx_wqe(&wqe->ctrl, &state->tls);
#endif
#ifdef CONFIG_MLX5_EN_IPSEC
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c
index 97cbea7ed048..82dc09aaa7fc 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c
@@ -298,12 +298,6 @@ bool mlx5e_tls_handle_tx_skb(struct net_device *netdev, struct mlx5e_txqsq *sq,
return false;
}
-void mlx5e_tls_handle_tx_wqe(struct mlx5e_txqsq *sq, struct mlx5_wqe_ctrl_seg *cseg,
- struct mlx5e_accel_tx_tls_state *state)
-{
- cseg->tis_tir_num = cpu_to_be32(state->tls_tisn << 8);
-}
-
static int tls_update_resync_sn(struct net_device *netdev,
struct sk_buff *skb,
struct mlx5e_tls_metadata *mdata)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h
index 5c3443200fd6..0ca0a023fb8d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h
@@ -53,8 +53,12 @@ static inline bool mlx5e_tls_skb_offloaded(struct sk_buff *skb)
return skb->sk && tls_is_sk_tx_device_offloaded(skb->sk);
}
-void mlx5e_tls_handle_tx_wqe(struct mlx5e_txqsq *sq, struct mlx5_wqe_ctrl_seg *cseg,
- struct mlx5e_accel_tx_tls_state *state);
+static inline void
+mlx5e_tls_handle_tx_wqe(struct mlx5_wqe_ctrl_seg *cseg,
+ struct mlx5e_accel_tx_tls_state *state)
+{
+ cseg->tis_tir_num = cpu_to_be32(state->tls_tisn << 8);
+}
void mlx5e_tls_handle_rx_skb_metadata(struct mlx5e_rq *rq, struct sk_buff *skb,
u32 *cqe_bcnt);
--
2.30.2
^ permalink raw reply related [flat|nested] 16+ messages in thread* [net-next 05/14] net/mlx5e: kTLS, Add resiliency to RX resync failures
2021-04-16 18:54 [pull request][net-next 00/14] mlx5 updates 2021-04-16 Saeed Mahameed
` (3 preceding siblings ...)
2021-04-16 18:54 ` [net-next 04/14] net/mlx5e: TX, Inline function mlx5e_tls_handle_tx_wqe() Saeed Mahameed
@ 2021-04-16 18:54 ` Saeed Mahameed
2021-04-16 18:54 ` [net-next 06/14] net/mlx5e: Allow mlx5e_safe_switch_channels to work with channels closed Saeed Mahameed
` (8 subsequent siblings)
13 siblings, 0 replies; 16+ messages in thread
From: Saeed Mahameed @ 2021-04-16 18:54 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski; +Cc: Tariq Toukan, netdev, Saeed Mahameed
From: Tariq Toukan <tariqt@nvidia.com>
When the TLS logic finds a tcp seq match for a kTLS RX resync
request, it calls the driver callback function mlx5e_ktls_resync()
to handle it and communicate it to the device.
Errors might occur during mlx5e_ktls_resync(), however, they are not
reported to the stack. Moreover, there is no error handling in the
stack for these errors.
In this patch, the driver obtains responsibility on errors handling,
adding queue and retry mechanisms to these resyncs.
We maintain a linked list of resync matches, and try posting them
to the async ICOSQ in the NAPI context.
Only possible failure that demands driver handling is ICOSQ being full.
By relying on the NAPI mechanism, we make sure that the entries in list
will be handled when ICOSQ completions arrive and make some room
available.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en.h | 4 +
.../ethernet/mellanox/mlx5/core/en/params.c | 3 +
.../ethernet/mellanox/mlx5/core/en/params.h | 1 +
.../mellanox/mlx5/core/en_accel/ktls.h | 11 ++
.../mellanox/mlx5/core/en_accel/ktls_rx.c | 124 +++++++++++++++---
.../mellanox/mlx5/core/en_accel/ktls_txrx.h | 20 +++
.../net/ethernet/mellanox/mlx5/core/en_main.c | 11 ++
.../ethernet/mellanox/mlx5/core/en_stats.c | 3 +
.../ethernet/mellanox/mlx5/core/en_stats.h | 2 +
.../net/ethernet/mellanox/mlx5/core/en_txrx.c | 5 +
10 files changed, 166 insertions(+), 18 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index cb4e7aaa4f8a..28a68eef8ae6 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -327,6 +327,7 @@ enum {
MLX5E_SQ_STATE_AM,
MLX5E_SQ_STATE_VLAN_NEED_L2_INLINE,
MLX5E_SQ_STATE_PENDING_XSK_TX,
+ MLX5E_SQ_STATE_PENDING_TLS_RX_RESYNC,
};
struct mlx5e_tx_mpwqe {
@@ -499,6 +500,8 @@ struct mlx5e_xdpsq {
struct mlx5e_channel *channel;
} ____cacheline_aligned_in_smp;
+struct mlx5e_ktls_resync_resp;
+
struct mlx5e_icosq {
/* data path */
u16 cc;
@@ -518,6 +521,7 @@ struct mlx5e_icosq {
u32 sqn;
u16 reserved_room;
unsigned long state;
+ struct mlx5e_ktls_resync_resp *ktls_resync;
/* control path */
struct mlx5_wq_ctrl wq_ctrl;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/params.c b/drivers/net/ethernet/mellanox/mlx5/core/en/params.c
index 7b2b52e75222..f6ba568e00be 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/params.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/params.c
@@ -621,6 +621,9 @@ static void mlx5e_build_async_icosq_param(struct mlx5_core_dev *mdev,
mlx5e_build_sq_param_common(mdev, param);
param->stop_room = mlx5e_stop_room_for_wqe(1); /* for XSK NOP */
+ param->is_tls = mlx5_accel_is_ktls_rx(mdev);
+ if (param->is_tls)
+ param->stop_room += mlx5e_stop_room_for_wqe(1); /* for TLS RX resync NOP */
MLX5_SET(sqc, sqc, reg_umr, MLX5_CAP_ETH(mdev, reg_umr_sq));
MLX5_SET(wq, wq, log_wq_sz, log_wq_size);
mlx5e_build_ico_cq_param(mdev, log_wq_size, ¶m->cqp);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/params.h b/drivers/net/ethernet/mellanox/mlx5/core/en/params.h
index 602e41a2bddd..fcc51ec6084e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/params.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/params.h
@@ -30,6 +30,7 @@ struct mlx5e_sq_param {
u32 sqc[MLX5_ST_SZ_DW(sqc)];
struct mlx5_wq_param wq;
bool is_mpw;
+ bool is_tls;
u16 stop_room;
};
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls.h
index baa58b62e8df..aaa579bf9a39 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls.h
@@ -12,6 +12,9 @@ void mlx5e_ktls_build_netdev(struct mlx5e_priv *priv);
int mlx5e_ktls_init_rx(struct mlx5e_priv *priv);
void mlx5e_ktls_cleanup_rx(struct mlx5e_priv *priv);
int mlx5e_ktls_set_feature_rx(struct net_device *netdev, bool enable);
+struct mlx5e_ktls_resync_resp *
+mlx5e_ktls_rx_resync_create_resp_list(void);
+void mlx5e_ktls_rx_resync_destroy_resp_list(struct mlx5e_ktls_resync_resp *resp_list);
#else
static inline void mlx5e_ktls_build_netdev(struct mlx5e_priv *priv)
@@ -33,6 +36,14 @@ static inline int mlx5e_ktls_set_feature_rx(struct net_device *netdev, bool enab
return -EOPNOTSUPP;
}
+static inline struct mlx5e_ktls_resync_resp *
+mlx5e_ktls_rx_resync_create_resp_list(void)
+{
+ return ERR_PTR(-EOPNOTSUPP);
+}
+
+static inline void
+mlx5e_ktls_rx_resync_destroy_resp_list(struct mlx5e_ktls_resync_resp *resp_list) {}
#endif
#endif /* __MLX5E_TLS_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c
index 76fd4b230003..4e58fade7a60 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c
@@ -56,6 +56,7 @@ struct mlx5e_ktls_offload_context_rx {
/* resync */
struct mlx5e_ktls_rx_resync_ctx resync;
+ struct list_head list;
};
static bool mlx5e_ktls_priv_rx_put(struct mlx5e_ktls_offload_context_rx *priv_rx)
@@ -72,6 +73,32 @@ static void mlx5e_ktls_priv_rx_get(struct mlx5e_ktls_offload_context_rx *priv_rx
refcount_inc(&priv_rx->resync.refcnt);
}
+struct mlx5e_ktls_resync_resp {
+ /* protects list changes */
+ spinlock_t lock;
+ struct list_head list;
+};
+
+void mlx5e_ktls_rx_resync_destroy_resp_list(struct mlx5e_ktls_resync_resp *resp_list)
+{
+ kvfree(resp_list);
+}
+
+struct mlx5e_ktls_resync_resp *
+mlx5e_ktls_rx_resync_create_resp_list(void)
+{
+ struct mlx5e_ktls_resync_resp *resp_list;
+
+ resp_list = kvzalloc(sizeof(*resp_list), GFP_KERNEL);
+ if (!resp_list)
+ return ERR_PTR(-ENOMEM);
+
+ INIT_LIST_HEAD(&resp_list->list);
+ spin_lock_init(&resp_list->lock);
+
+ return resp_list;
+}
+
static int mlx5e_ktls_create_tir(struct mlx5_core_dev *mdev, u32 *tirn, u32 rqtn)
{
int err, inlen;
@@ -358,33 +385,32 @@ static void resync_init(struct mlx5e_ktls_rx_resync_ctx *resync,
/* Function can be called with the refcount being either elevated or not.
* It does not affect the refcount.
*/
-static int resync_handle_seq_match(struct mlx5e_ktls_offload_context_rx *priv_rx,
- struct mlx5e_channel *c)
+static void resync_handle_seq_match(struct mlx5e_ktls_offload_context_rx *priv_rx,
+ struct mlx5e_channel *c)
{
struct tls12_crypto_info_aes_gcm_128 *info = &priv_rx->crypto_info;
- struct mlx5_wqe_ctrl_seg *cseg;
+ struct mlx5e_ktls_resync_resp *ktls_resync;
struct mlx5e_icosq *sq;
- int err;
+ bool trigger_poll;
memcpy(info->rec_seq, &priv_rx->resync.sw_rcd_sn_be, sizeof(info->rec_seq));
- err = 0;
sq = &c->async_icosq;
- spin_lock_bh(&c->async_icosq_lock);
+ ktls_resync = sq->ktls_resync;
- cseg = post_static_params(sq, priv_rx);
- if (IS_ERR(cseg)) {
- priv_rx->rq_stats->tls_resync_res_skip++;
- err = PTR_ERR(cseg);
- goto unlock;
- }
- /* Do not increment priv_rx refcnt, CQE handling is empty */
- mlx5e_notify_hw(&sq->wq, sq->pc, sq->uar_map, cseg);
- priv_rx->rq_stats->tls_resync_res_ok++;
-unlock:
- spin_unlock_bh(&c->async_icosq_lock);
+ spin_lock_bh(&ktls_resync->lock);
+ list_add_tail(&priv_rx->list, &ktls_resync->list);
+ trigger_poll = !test_and_set_bit(MLX5E_SQ_STATE_PENDING_TLS_RX_RESYNC, &sq->state);
+ spin_unlock_bh(&ktls_resync->lock);
- return err;
+ if (!trigger_poll)
+ return;
+
+ if (!napi_if_scheduled_mark_missed(&c->napi)) {
+ spin_lock_bh(&c->async_icosq_lock);
+ mlx5e_trigger_irq(sq);
+ spin_unlock_bh(&c->async_icosq_lock);
+ }
}
/* Function can be called with the refcount being either elevated or not.
@@ -675,3 +701,65 @@ void mlx5e_ktls_del_rx(struct net_device *netdev, struct tls_context *tls_ctx)
*/
mlx5e_ktls_priv_rx_put(priv_rx);
}
+
+bool mlx5e_ktls_rx_handle_resync_list(struct mlx5e_channel *c, int budget)
+{
+ struct mlx5e_ktls_offload_context_rx *priv_rx, *tmp;
+ struct mlx5e_ktls_resync_resp *ktls_resync;
+ struct mlx5_wqe_ctrl_seg *db_cseg;
+ struct mlx5e_icosq *sq;
+ LIST_HEAD(local_list);
+ int i, j;
+
+ sq = &c->async_icosq;
+
+ if (unlikely(!test_bit(MLX5E_SQ_STATE_ENABLED, &sq->state)))
+ return false;
+
+ ktls_resync = sq->ktls_resync;
+ db_cseg = NULL;
+ i = 0;
+
+ spin_lock(&ktls_resync->lock);
+ list_for_each_entry_safe(priv_rx, tmp, &ktls_resync->list, list) {
+ list_move(&priv_rx->list, &local_list);
+ if (++i == budget)
+ break;
+ }
+ if (list_empty(&ktls_resync->list))
+ clear_bit(MLX5E_SQ_STATE_PENDING_TLS_RX_RESYNC, &sq->state);
+ spin_unlock(&ktls_resync->lock);
+
+ spin_lock(&c->async_icosq_lock);
+ for (j = 0; j < i; j++) {
+ struct mlx5_wqe_ctrl_seg *cseg;
+
+ priv_rx = list_first_entry(&local_list,
+ struct mlx5e_ktls_offload_context_rx,
+ list);
+ cseg = post_static_params(sq, priv_rx);
+ if (IS_ERR(cseg))
+ break;
+ list_del(&priv_rx->list);
+ db_cseg = cseg;
+ }
+ if (db_cseg)
+ mlx5e_notify_hw(&sq->wq, sq->pc, sq->uar_map, db_cseg);
+ spin_unlock(&c->async_icosq_lock);
+
+ priv_rx->rq_stats->tls_resync_res_ok += j;
+
+ if (!list_empty(&local_list)) {
+ /* This happens only if ICOSQ is full.
+ * There is no need to mark busy or explicitly ask for a NAPI cycle,
+ * it will be triggered by the outstanding ICOSQ completions.
+ */
+ spin_lock(&ktls_resync->lock);
+ list_splice(&local_list, &ktls_resync->list);
+ set_bit(MLX5E_SQ_STATE_PENDING_TLS_RX_RESYNC, &sq->state);
+ spin_unlock(&ktls_resync->lock);
+ priv_rx->rq_stats->tls_resync_res_retry++;
+ }
+
+ return i == budget;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_txrx.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_txrx.h
index ee04e916fa21..8f79335057dc 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_txrx.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_txrx.h
@@ -40,6 +40,14 @@ mlx5e_ktls_tx_try_handle_resync_dump_comp(struct mlx5e_txqsq *sq,
}
return false;
}
+
+bool mlx5e_ktls_rx_handle_resync_list(struct mlx5e_channel *c, int budget);
+
+static inline bool
+mlx5e_ktls_rx_pending_resync_list(struct mlx5e_channel *c, int budget)
+{
+ return budget && test_bit(MLX5E_SQ_STATE_PENDING_TLS_RX_RESYNC, &c->async_icosq.state);
+}
#else
static inline bool
mlx5e_ktls_tx_try_handle_resync_dump_comp(struct mlx5e_txqsq *sq,
@@ -49,6 +57,18 @@ mlx5e_ktls_tx_try_handle_resync_dump_comp(struct mlx5e_txqsq *sq,
return false;
}
+static inline bool
+mlx5e_ktls_rx_handle_resync_list(struct mlx5e_channel *c, int budget)
+{
+ return false;
+}
+
+static inline bool
+mlx5e_ktls_rx_pending_resync_list(struct mlx5e_channel *c, int budget)
+{
+ return false;
+}
+
#endif /* CONFIG_MLX5_EN_TLS */
#endif /* __MLX5E_TLS_TXRX_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 64d6c0fd92bf..df4959e46f27 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -1409,8 +1409,17 @@ int mlx5e_open_icosq(struct mlx5e_channel *c, struct mlx5e_params *params,
if (err)
goto err_free_icosq;
+ if (param->is_tls) {
+ sq->ktls_resync = mlx5e_ktls_rx_resync_create_resp_list();
+ if (IS_ERR(sq->ktls_resync)) {
+ err = PTR_ERR(sq->ktls_resync);
+ goto err_destroy_icosq;
+ }
+ }
return 0;
+err_destroy_icosq:
+ mlx5e_destroy_sq(c->mdev, sq->sqn);
err_free_icosq:
mlx5e_free_icosq(sq);
@@ -1432,6 +1441,8 @@ void mlx5e_close_icosq(struct mlx5e_icosq *sq)
{
struct mlx5e_channel *c = sq->channel;
+ if (sq->ktls_resync)
+ mlx5e_ktls_rx_resync_destroy_resp_list(sq->ktls_resync);
mlx5e_destroy_sq(c->mdev, sq->sqn);
mlx5e_free_icosq_descs(sq);
mlx5e_free_icosq(sq);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
index 353513bd0d5e..5146aa200de9 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
@@ -184,6 +184,7 @@ static const struct counter_desc sw_stats_desc[] = {
{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_tls_resync_req_end) },
{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_tls_resync_req_skip) },
{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_tls_resync_res_ok) },
+ { MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_tls_resync_res_retry) },
{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_tls_resync_res_skip) },
{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_tls_err) },
#endif
@@ -344,6 +345,7 @@ static void mlx5e_stats_grp_sw_update_stats_rq_stats(struct mlx5e_sw_stats *s,
s->rx_tls_resync_req_end += rq_stats->tls_resync_req_end;
s->rx_tls_resync_req_skip += rq_stats->tls_resync_req_skip;
s->rx_tls_resync_res_ok += rq_stats->tls_resync_res_ok;
+ s->rx_tls_resync_res_retry += rq_stats->tls_resync_res_retry;
s->rx_tls_resync_res_skip += rq_stats->tls_resync_res_skip;
s->rx_tls_err += rq_stats->tls_err;
#endif
@@ -1654,6 +1656,7 @@ static const struct counter_desc rq_stats_desc[] = {
{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, tls_resync_req_end) },
{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, tls_resync_req_skip) },
{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, tls_resync_res_ok) },
+ { MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, tls_resync_res_retry) },
{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, tls_resync_res_skip) },
{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, tls_err) },
#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h
index 3f0789e51eed..9614de49b7ac 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h
@@ -209,6 +209,7 @@ struct mlx5e_sw_stats {
u64 rx_tls_resync_req_end;
u64 rx_tls_resync_req_skip;
u64 rx_tls_resync_res_ok;
+ u64 rx_tls_resync_res_retry;
u64 rx_tls_resync_res_skip;
u64 rx_tls_err;
#endif
@@ -339,6 +340,7 @@ struct mlx5e_rq_stats {
u64 tls_resync_req_end;
u64 tls_resync_req_skip;
u64 tls_resync_res_ok;
+ u64 tls_resync_res_retry;
u64 tls_resync_res_skip;
u64 tls_err;
#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
index d54da3797c30..833be29170a1 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
@@ -36,6 +36,7 @@
#include "en/xdp.h"
#include "en/xsk/rx.h"
#include "en/xsk/tx.h"
+#include "en_accel/ktls_txrx.h"
static inline bool mlx5e_channel_no_affinity_change(struct mlx5e_channel *c)
{
@@ -171,6 +172,10 @@ int mlx5e_napi_poll(struct napi_struct *napi, int budget)
*/
clear_bit(MLX5E_SQ_STATE_PENDING_XSK_TX, &c->async_icosq.state);
+ /* Keep after async ICOSQ CQ poll */
+ if (unlikely(mlx5e_ktls_rx_pending_resync_list(c, budget)))
+ busy |= mlx5e_ktls_rx_handle_resync_list(c, budget);
+
busy |= INDIRECT_CALL_2(rq->post_wqes,
mlx5e_post_rx_mpwqes,
mlx5e_post_rx_wqes,
--
2.30.2
^ permalink raw reply related [flat|nested] 16+ messages in thread* [net-next 06/14] net/mlx5e: Allow mlx5e_safe_switch_channels to work with channels closed
2021-04-16 18:54 [pull request][net-next 00/14] mlx5 updates 2021-04-16 Saeed Mahameed
` (4 preceding siblings ...)
2021-04-16 18:54 ` [net-next 05/14] net/mlx5e: kTLS, Add resiliency to RX resync failures Saeed Mahameed
@ 2021-04-16 18:54 ` Saeed Mahameed
2021-04-16 18:54 ` [net-next 07/14] net/mlx5e: Use mlx5e_safe_switch_channels when channels are closed Saeed Mahameed
` (7 subsequent siblings)
13 siblings, 0 replies; 16+ messages in thread
From: Saeed Mahameed @ 2021-04-16 18:54 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: Tariq Toukan, netdev, Maxim Mikityanskiy, Saeed Mahameed
From: Maxim Mikityanskiy <maximmi@mellanox.com>
mlx5e_safe_switch_channels is used to modify channel parameters and/or
hardware configuration in a safe way, so that if anything goes wrong,
everything reverts to the old configuration and remains in a consistent
state.
However, this function only works when the channels are open. When the
caller needs to modify some parameters, first it has to check that the
channels are open, otherwise it has to assign parameters directly, and
such boilerplate repeats in many different places.
This commit prepares for the refactoring of such places by allowing
mlx5e_safe_switch_channels to work when the channels are closed. In this
case it will assign the new parameters and run the preactivate hook.
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../net/ethernet/mellanox/mlx5/core/en_main.c | 40 +++++++++++++------
1 file changed, 27 insertions(+), 13 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index df4959e46f27..cb88d7239db6 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -2852,12 +2852,16 @@ static int mlx5e_switch_priv_channels(struct mlx5e_priv *priv,
struct net_device *netdev = priv->netdev;
struct mlx5e_channels old_chs;
int carrier_ok;
+ bool opened;
int err = 0;
- carrier_ok = netif_carrier_ok(netdev);
- netif_carrier_off(netdev);
+ opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
+ if (opened) {
+ carrier_ok = netif_carrier_ok(netdev);
+ netif_carrier_off(netdev);
- mlx5e_deactivate_priv_channels(priv);
+ mlx5e_deactivate_priv_channels(priv);
+ }
old_chs = priv->channels;
priv->channels = *new_chs;
@@ -2873,15 +2877,19 @@ static int mlx5e_switch_priv_channels(struct mlx5e_priv *priv,
}
}
- mlx5e_close_channels(&old_chs);
- priv->profile->update_rx(priv);
+ if (opened) {
+ mlx5e_close_channels(&old_chs);
+ priv->profile->update_rx(priv);
+ }
out:
- mlx5e_activate_priv_channels(priv);
+ if (opened) {
+ mlx5e_activate_priv_channels(priv);
- /* return carrier back if needed */
- if (carrier_ok)
- netif_carrier_on(netdev);
+ /* return carrier back if needed */
+ if (carrier_ok)
+ netif_carrier_on(netdev);
+ }
return err;
}
@@ -2891,11 +2899,16 @@ int mlx5e_safe_switch_channels(struct mlx5e_priv *priv,
mlx5e_fp_preactivate preactivate,
void *context)
{
+ bool opened;
int err;
- err = mlx5e_open_channels(priv, new_chs);
- if (err)
- return err;
+ opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
+
+ if (opened) {
+ err = mlx5e_open_channels(priv, new_chs);
+ if (err)
+ return err;
+ }
err = mlx5e_switch_priv_channels(priv, new_chs, preactivate, context);
if (err)
@@ -2904,7 +2917,8 @@ int mlx5e_safe_switch_channels(struct mlx5e_priv *priv,
return 0;
err_close:
- mlx5e_close_channels(new_chs);
+ if (opened)
+ mlx5e_close_channels(new_chs);
return err;
}
--
2.30.2
^ permalink raw reply related [flat|nested] 16+ messages in thread* [net-next 07/14] net/mlx5e: Use mlx5e_safe_switch_channels when channels are closed
2021-04-16 18:54 [pull request][net-next 00/14] mlx5 updates 2021-04-16 Saeed Mahameed
` (5 preceding siblings ...)
2021-04-16 18:54 ` [net-next 06/14] net/mlx5e: Allow mlx5e_safe_switch_channels to work with channels closed Saeed Mahameed
@ 2021-04-16 18:54 ` Saeed Mahameed
2021-04-16 18:54 ` [net-next 08/14] net/mlx5e: Refactor on-the-fly configuration changes Saeed Mahameed
` (6 subsequent siblings)
13 siblings, 0 replies; 16+ messages in thread
From: Saeed Mahameed @ 2021-04-16 18:54 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: Tariq Toukan, netdev, Maxim Mikityanskiy, Saeed Mahameed
From: Maxim Mikityanskiy <maximmi@mellanox.com>
This commit uses new functionality of mlx5e_safe_switch_channels
introduced by the previous commit to reduce the amount of repeating
similar code all over the driver.
It's very common in mlx5e to call mlx5e_safe_switch_channels when the
channels are open, but assign parameters and run hardware commands
manually when the channels are closed.
After the previous commit it's no longer needed to do such manual things
every time, so this commit removes unneeded code and relies on the new
functionality of mlx5e_safe_switch_channels. Some of the places are
refactored and simplified, where more complex flows are used to change
configuration on the fly, without recreating the channels (the logic is
rewritten in a more robust way, with a reset required by default and a
list of exceptions).
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../net/ethernet/mellanox/mlx5/core/en/ptp.c | 3 +
.../ethernet/mellanox/mlx5/core/en_dcbnl.c | 23 ++-----
.../ethernet/mellanox/mlx5/core/en_ethtool.c | 65 +++----------------
.../net/ethernet/mellanox/mlx5/core/en_main.c | 65 ++++++-------------
.../ethernet/mellanox/mlx5/core/ipoib/ipoib.c | 11 +---
5 files changed, 40 insertions(+), 127 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/ptp.c b/drivers/net/ethernet/mellanox/mlx5/core/en/ptp.c
index 72e7dd6d78c0..d907c1acd4d5 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/ptp.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/ptp.c
@@ -792,6 +792,9 @@ int mlx5e_ptp_rx_manage_fs(struct mlx5e_priv *priv, bool set)
if (!priv->profile->rx_ptp_support)
return 0;
+ if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+ return 0;
+
if (set) {
if (!c || !test_bit(MLX5E_PTP_STATE_RX, c->state)) {
netdev_WARN_ONCE(priv->netdev, "Don't try to add PTP RX-FS rules");
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
index f23c67575073..9d4d83159603 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
@@ -1150,8 +1150,6 @@ static int mlx5e_update_trust_state_hw(struct mlx5e_priv *priv, void *context)
static int mlx5e_set_trust_state(struct mlx5e_priv *priv, u8 trust_state)
{
struct mlx5e_channels new_channels = {};
- bool reset_channels = true;
- bool opened;
int err = 0;
mutex_lock(&priv->state_lock);
@@ -1160,25 +1158,18 @@ static int mlx5e_set_trust_state(struct mlx5e_priv *priv, u8 trust_state)
mlx5e_params_calc_trust_tx_min_inline_mode(priv->mdev, &new_channels.params,
trust_state);
- opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
- if (!opened)
- reset_channels = false;
-
/* Skip if tx_min_inline is the same */
if (new_channels.params.tx_min_inline_mode ==
- priv->channels.params.tx_min_inline_mode)
- reset_channels = false;
-
- if (reset_channels) {
- err = mlx5e_safe_switch_channels(priv, &new_channels,
- mlx5e_update_trust_state_hw,
- &trust_state);
- } else {
+ priv->channels.params.tx_min_inline_mode) {
err = mlx5e_update_trust_state_hw(priv, &trust_state);
- if (!err && !opened)
- priv->channels.params = new_channels.params;
+ goto out;
}
+ err = mlx5e_safe_switch_channels(priv, &new_channels,
+ mlx5e_update_trust_state_hw,
+ &trust_state);
+
+out:
mutex_unlock(&priv->state_lock);
return err;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
index f17690cbeeea..6a15666f106f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -373,11 +373,6 @@ int mlx5e_ethtool_set_ringparam(struct mlx5e_priv *priv,
if (err)
goto unlock;
- if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
- priv->channels.params = new_channels.params;
- goto unlock;
- }
-
err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
unlock:
@@ -425,6 +420,7 @@ int mlx5e_ethtool_set_channels(struct mlx5e_priv *priv,
unsigned int count = ch->combined_count;
struct mlx5e_channels new_channels = {};
bool arfs_enabled;
+ bool opened;
int err = 0;
if (!count) {
@@ -462,19 +458,9 @@ int mlx5e_ethtool_set_channels(struct mlx5e_priv *priv,
new_channels.params = *cur_params;
new_channels.params.num_channels = count;
- if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
- struct mlx5e_params old_params;
-
- old_params = *cur_params;
- *cur_params = new_channels.params;
- err = mlx5e_num_channels_changed(priv);
- if (err)
- *cur_params = old_params;
+ opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
- goto out;
- }
-
- arfs_enabled = priv->netdev->features & NETIF_F_NTUPLE;
+ arfs_enabled = opened && (priv->netdev->features & NETIF_F_NTUPLE);
if (arfs_enabled)
mlx5e_arfs_disable(priv);
@@ -625,12 +611,10 @@ int mlx5e_ethtool_set_coalesce(struct mlx5e_priv *priv,
mlx5e_reset_tx_moderation(&new_channels.params, mode);
}
- if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
- priv->channels.params = new_channels.params;
- goto out;
- }
-
- if (!reset_rx && !reset_tx) {
+ /* If DIM state hasn't changed, it's possible to modify interrupt
+ * moderation parameters on the fly, even if the channels are open.
+ */
+ if (!reset_rx && !reset_tx && test_bit(MLX5E_STATE_OPENED, &priv->state)) {
if (!coal->use_adaptive_rx_coalesce)
mlx5e_set_priv_channels_rx_coalesce(priv, coal);
if (!coal->use_adaptive_tx_coalesce)
@@ -1885,11 +1869,6 @@ static int set_pflag_cqe_based_moder(struct net_device *netdev, bool enable,
else
mlx5e_set_tx_cq_mode_params(&new_channels.params, cq_period_mode);
- if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
- priv->channels.params = new_channels.params;
- return 0;
- }
-
return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
}
@@ -1920,10 +1899,6 @@ int mlx5e_modify_rx_cqe_compression_locked(struct mlx5e_priv *priv, bool new_val
if (priv->tstamp.rx_filter != HWTSTAMP_FILTER_NONE)
new_channels.params.ptp_rx = new_val;
- if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
- priv->channels.params = new_channels.params;
- return 0;
- }
if (new_channels.params.ptp_rx == priv->channels.params.ptp_rx)
err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
@@ -1980,11 +1955,6 @@ static int set_pflag_rx_striding_rq(struct net_device *netdev, bool enable)
MLX5E_SET_PFLAG(&new_channels.params, MLX5E_PFLAG_RX_STRIDING_RQ, enable);
mlx5e_set_rq_type(mdev, &new_channels.params);
- if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
- priv->channels.params = new_channels.params;
- return 0;
- }
-
return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
}
@@ -2015,7 +1985,6 @@ static int set_pflag_tx_mpwqe_common(struct net_device *netdev, u32 flag, bool e
struct mlx5e_priv *priv = netdev_priv(netdev);
struct mlx5_core_dev *mdev = priv->mdev;
struct mlx5e_channels new_channels = {};
- int err;
if (enable && !MLX5_CAP_ETH(mdev, enhanced_multi_pkt_send_wqe))
return -EOPNOTSUPP;
@@ -2024,13 +1993,7 @@ static int set_pflag_tx_mpwqe_common(struct net_device *netdev, u32 flag, bool e
MLX5E_SET_PFLAG(&new_channels.params, flag, enable);
- if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
- priv->channels.params = new_channels.params;
- return 0;
- }
-
- err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
- return err;
+ return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
}
static int set_pflag_xdp_tx_mpwqe(struct net_device *netdev, bool enable)
@@ -2070,20 +2033,8 @@ static int set_pflag_tx_port_ts(struct net_device *netdev, bool enable)
* has the same log_sq_size.
*/
- if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
- struct mlx5e_params old_params;
-
- old_params = priv->channels.params;
- priv->channels.params = new_channels.params;
- err = mlx5e_num_channels_changed(priv);
- if (err)
- priv->channels.params = old_params;
- goto out;
- }
-
err = mlx5e_safe_switch_channels(priv, &new_channels,
mlx5e_num_channels_changed_ctx, NULL);
-out:
if (!err)
priv->tx_ptp_opened = true;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index cb88d7239db6..0e5539afc3a0 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -3407,18 +3407,6 @@ static int mlx5e_setup_tc_mqprio(struct mlx5e_priv *priv,
new_channels.params = priv->channels.params;
new_channels.params.num_tc = tc ? tc : 1;
- if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
- struct mlx5e_params old_params;
-
- old_params = priv->channels.params;
- priv->channels.params = new_channels.params;
- err = mlx5e_num_channels_changed(priv);
- if (err)
- priv->channels.params = old_params;
-
- goto out;
- }
-
err = mlx5e_safe_switch_channels(priv, &new_channels,
mlx5e_num_channels_changed_ctx, NULL);
@@ -3647,8 +3635,8 @@ static int set_feature_lro(struct net_device *netdev, bool enable)
struct mlx5_core_dev *mdev = priv->mdev;
struct mlx5e_channels new_channels = {};
struct mlx5e_params *cur_params;
+ bool skip_reset = false;
int err = 0;
- bool reset;
mutex_lock(&priv->state_lock);
@@ -3666,18 +3654,16 @@ static int set_feature_lro(struct net_device *netdev, bool enable)
goto out;
}
- reset = test_bit(MLX5E_STATE_OPENED, &priv->state);
-
new_channels.params = *cur_params;
new_channels.params.lro_en = enable;
- if (cur_params->rq_wq_type != MLX5_WQ_TYPE_CYCLIC) {
+ if (cur_params->rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ) {
if (mlx5e_rx_mpwqe_is_linear_skb(mdev, cur_params, NULL) ==
mlx5e_rx_mpwqe_is_linear_skb(mdev, &new_channels.params, NULL))
- reset = false;
+ skip_reset = true;
}
- if (!reset) {
+ if (skip_reset) {
struct mlx5e_params old_params;
old_params = *cur_params;
@@ -3920,16 +3906,13 @@ int mlx5e_change_mtu(struct net_device *netdev, int new_mtu,
struct mlx5e_priv *priv = netdev_priv(netdev);
struct mlx5e_channels new_channels = {};
struct mlx5e_params *params;
+ bool skip_reset = false;
int err = 0;
- bool reset;
mutex_lock(&priv->state_lock);
params = &priv->channels.params;
- reset = !params->lro_en;
- reset = reset && test_bit(MLX5E_STATE_OPENED, &priv->state);
-
new_channels.params = *params;
new_channels.params.sw_mtu = new_mtu;
err = mlx5e_validate_params(priv->mdev, &new_channels.params);
@@ -3951,21 +3934,26 @@ int mlx5e_change_mtu(struct net_device *netdev, int new_mtu,
goto out;
}
+ if (params->lro_en)
+ skip_reset = true;
+
if (params->rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ) {
- bool is_linear = mlx5e_rx_mpwqe_is_linear_skb(priv->mdev,
- &new_channels.params,
- NULL);
+ bool is_linear_old = mlx5e_rx_mpwqe_is_linear_skb(priv->mdev, params, NULL);
+ bool is_linear_new = mlx5e_rx_mpwqe_is_linear_skb(priv->mdev,
+ &new_channels.params, NULL);
u8 ppw_old = mlx5e_mpwqe_log_pkts_per_wqe(params, NULL);
u8 ppw_new = mlx5e_mpwqe_log_pkts_per_wqe(&new_channels.params, NULL);
- /* If XSK is active, XSK RQs are linear. */
- is_linear |= priv->xsk.refcnt;
-
- /* Always reset in linear mode - hw_mtu is used in data path. */
- reset = reset && (is_linear || (ppw_old != ppw_new));
+ /* Always reset in linear mode - hw_mtu is used in data path.
+ * Check that the mode was non-linear and didn't change.
+ * If XSK is active, XSK RQs are linear.
+ */
+ if (!is_linear_old && !is_linear_new && !priv->xsk.refcnt &&
+ ppw_old == ppw_new)
+ skip_reset = true;
}
- if (!reset) {
+ if (skip_reset) {
unsigned int old_mtu = params->sw_mtu;
params->sw_mtu = new_mtu;
@@ -3976,17 +3964,13 @@ int mlx5e_change_mtu(struct net_device *netdev, int new_mtu,
goto out;
}
}
- netdev->mtu = params->sw_mtu;
goto out;
}
err = mlx5e_safe_switch_channels(priv, &new_channels, preactivate, NULL);
- if (err)
- goto out;
-
- netdev->mtu = new_channels.params.sw_mtu;
out:
+ netdev->mtu = params->sw_mtu;
mutex_unlock(&priv->state_lock);
return err;
}
@@ -4061,10 +4045,6 @@ int mlx5e_hwstamp_set(struct mlx5e_priv *priv, struct ifreq *ifr)
if (new_channels.params.ptp_rx == priv->channels.params.ptp_rx)
goto out;
- if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
- priv->channels.params = new_channels.params;
- goto out;
- }
err = mlx5e_safe_switch_channels(priv, &new_channels, mlx5e_ptp_rx_manage_fs_ctx,
&new_channels.params.ptp_rx);
if (err) {
@@ -4449,7 +4429,7 @@ static int mlx5e_xdp_set(struct net_device *netdev, struct bpf_prog *prog)
*/
bpf_prog_add(prog, priv->channels.num);
- if (was_opened && reset) {
+ if (reset) {
struct mlx5e_channels new_channels = {};
new_channels.params = priv->channels.params;
@@ -4470,9 +4450,6 @@ static int mlx5e_xdp_set(struct net_device *netdev, struct bpf_prog *prog)
if (old_prog)
bpf_prog_put(old_prog);
- if (!was_opened && reset) /* change RQ type according to priv->xdp_prog */
- mlx5e_set_rq_type(priv->mdev, &priv->channels.params);
-
if (!was_opened || reset)
goto unlock;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
index b65b0cefc5b3..9555127ce7e7 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
@@ -482,20 +482,11 @@ static int mlx5i_change_mtu(struct net_device *netdev, int new_mtu)
{
struct mlx5e_priv *priv = mlx5i_epriv(netdev);
struct mlx5e_channels new_channels = {};
- struct mlx5e_params *params;
int err = 0;
mutex_lock(&priv->state_lock);
- params = &priv->channels.params;
-
- if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
- params->sw_mtu = new_mtu;
- netdev->mtu = params->sw_mtu;
- goto out;
- }
-
- new_channels.params = *params;
+ new_channels.params = priv->channels.params;
new_channels.params.sw_mtu = new_mtu;
err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
--
2.30.2
^ permalink raw reply related [flat|nested] 16+ messages in thread* [net-next 08/14] net/mlx5e: Refactor on-the-fly configuration changes
2021-04-16 18:54 [pull request][net-next 00/14] mlx5 updates 2021-04-16 Saeed Mahameed
` (6 preceding siblings ...)
2021-04-16 18:54 ` [net-next 07/14] net/mlx5e: Use mlx5e_safe_switch_channels when channels are closed Saeed Mahameed
@ 2021-04-16 18:54 ` Saeed Mahameed
2021-04-16 18:54 ` [net-next 09/14] net/mlx5e: Cleanup safe switch channels API by passing params Saeed Mahameed
` (5 subsequent siblings)
13 siblings, 0 replies; 16+ messages in thread
From: Saeed Mahameed @ 2021-04-16 18:54 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: Tariq Toukan, netdev, Maxim Mikityanskiy, Saeed Mahameed
From: Maxim Mikityanskiy <maximmi@mellanox.com>
This commit extends mlx5e_safe_switch_channels() to support on-the-fly
configuration changes, when the channels are open, but don't need to be
recreated. Such flows exist when a parameter being changed doesn't
affect how the queues are created, or when the queues can be modified
while remaining active.
Before this commit, such flows were handled as special cases on the
caller site. This commit adds this functionality to
mlx5e_safe_switch_channels(), allowing the caller to pass a boolean
indicating whether it's required to recreate the channels or it's
allowed to skip it. The logic of switching channel parameters is now
completely encapsulated into mlx5e_safe_switch_channels().
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en.h | 2 +-
.../ethernet/mellanox/mlx5/core/en_dcbnl.c | 12 +-
.../ethernet/mellanox/mlx5/core/en_ethtool.c | 24 ++-
.../net/ethernet/mellanox/mlx5/core/en_main.c | 151 +++++++-----------
.../ethernet/mellanox/mlx5/core/ipoib/ipoib.c | 2 +-
5 files changed, 78 insertions(+), 113 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index 28a68eef8ae6..d9ed20a4db53 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -1021,7 +1021,7 @@ int mlx5e_safe_reopen_channels(struct mlx5e_priv *priv);
int mlx5e_safe_switch_channels(struct mlx5e_priv *priv,
struct mlx5e_channels *new_chs,
mlx5e_fp_preactivate preactivate,
- void *context);
+ void *context, bool reset);
int mlx5e_update_tx_netdev_queues(struct mlx5e_priv *priv);
int mlx5e_num_channels_changed(struct mlx5e_priv *priv);
int mlx5e_num_channels_changed_ctx(struct mlx5e_priv *priv, void *context);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
index 9d4d83159603..0b0273fac6e3 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
@@ -1150,7 +1150,8 @@ static int mlx5e_update_trust_state_hw(struct mlx5e_priv *priv, void *context)
static int mlx5e_set_trust_state(struct mlx5e_priv *priv, u8 trust_state)
{
struct mlx5e_channels new_channels = {};
- int err = 0;
+ bool reset = true;
+ int err;
mutex_lock(&priv->state_lock);
@@ -1160,16 +1161,13 @@ static int mlx5e_set_trust_state(struct mlx5e_priv *priv, u8 trust_state)
/* Skip if tx_min_inline is the same */
if (new_channels.params.tx_min_inline_mode ==
- priv->channels.params.tx_min_inline_mode) {
- err = mlx5e_update_trust_state_hw(priv, &trust_state);
- goto out;
- }
+ priv->channels.params.tx_min_inline_mode)
+ reset = false;
err = mlx5e_safe_switch_channels(priv, &new_channels,
mlx5e_update_trust_state_hw,
- &trust_state);
+ &trust_state, reset);
-out:
mutex_unlock(&priv->state_lock);
return err;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
index 6a15666f106f..82994175aded 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -373,7 +373,7 @@ int mlx5e_ethtool_set_ringparam(struct mlx5e_priv *priv,
if (err)
goto unlock;
- err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
+ err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, true);
unlock:
mutex_unlock(&priv->state_lock);
@@ -466,7 +466,7 @@ int mlx5e_ethtool_set_channels(struct mlx5e_priv *priv,
/* Switch to new channels, set new parameters and close old ones */
err = mlx5e_safe_switch_channels(priv, &new_channels,
- mlx5e_num_channels_changed_ctx, NULL);
+ mlx5e_num_channels_changed_ctx, NULL, true);
if (arfs_enabled) {
int err2 = mlx5e_arfs_enable(priv);
@@ -563,6 +563,7 @@ int mlx5e_ethtool_set_coalesce(struct mlx5e_priv *priv,
struct mlx5_core_dev *mdev = priv->mdev;
struct mlx5e_channels new_channels = {};
bool reset_rx, reset_tx;
+ bool reset = true;
int err = 0;
if (!MLX5_CAP_GEN(mdev, cq_moderation))
@@ -619,13 +620,11 @@ int mlx5e_ethtool_set_coalesce(struct mlx5e_priv *priv,
mlx5e_set_priv_channels_rx_coalesce(priv, coal);
if (!coal->use_adaptive_tx_coalesce)
mlx5e_set_priv_channels_tx_coalesce(priv, coal);
- priv->channels.params = new_channels.params;
- goto out;
+ reset = false;
}
- err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
+ err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, reset);
-out:
mutex_unlock(&priv->state_lock);
return err;
}
@@ -1869,7 +1868,7 @@ static int set_pflag_cqe_based_moder(struct net_device *netdev, bool enable,
else
mlx5e_set_tx_cq_mode_params(&new_channels.params, cq_period_mode);
- return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
+ return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, true);
}
static int set_pflag_tx_cqe_based_moder(struct net_device *netdev, bool enable)
@@ -1899,12 +1898,11 @@ int mlx5e_modify_rx_cqe_compression_locked(struct mlx5e_priv *priv, bool new_val
if (priv->tstamp.rx_filter != HWTSTAMP_FILTER_NONE)
new_channels.params.ptp_rx = new_val;
-
if (new_channels.params.ptp_rx == priv->channels.params.ptp_rx)
- err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
+ err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, true);
else
err = mlx5e_safe_switch_channels(priv, &new_channels, mlx5e_ptp_rx_manage_fs_ctx,
- &new_channels.params.ptp_rx);
+ &new_channels.params.ptp_rx, true);
if (err)
return err;
@@ -1955,7 +1953,7 @@ static int set_pflag_rx_striding_rq(struct net_device *netdev, bool enable)
MLX5E_SET_PFLAG(&new_channels.params, MLX5E_PFLAG_RX_STRIDING_RQ, enable);
mlx5e_set_rq_type(mdev, &new_channels.params);
- return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
+ return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, true);
}
static int set_pflag_rx_no_csum_complete(struct net_device *netdev, bool enable)
@@ -1993,7 +1991,7 @@ static int set_pflag_tx_mpwqe_common(struct net_device *netdev, u32 flag, bool e
MLX5E_SET_PFLAG(&new_channels.params, flag, enable);
- return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
+ return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, true);
}
static int set_pflag_xdp_tx_mpwqe(struct net_device *netdev, bool enable)
@@ -2034,7 +2032,7 @@ static int set_pflag_tx_port_ts(struct net_device *netdev, bool enable)
*/
err = mlx5e_safe_switch_channels(priv, &new_channels,
- mlx5e_num_channels_changed_ctx, NULL);
+ mlx5e_num_channels_changed_ctx, NULL, true);
if (!err)
priv->tx_ptp_opened = true;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 0e5539afc3a0..7686d4997696 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -2844,6 +2844,29 @@ void mlx5e_deactivate_priv_channels(struct mlx5e_priv *priv)
mlx5e_deactivate_channels(&priv->channels);
}
+static int mlx5e_switch_priv_params(struct mlx5e_priv *priv,
+ struct mlx5e_params *new_params,
+ mlx5e_fp_preactivate preactivate,
+ void *context)
+{
+ struct mlx5e_params old_params;
+
+ old_params = priv->channels.params;
+ priv->channels.params = *new_params;
+
+ if (preactivate) {
+ int err;
+
+ err = preactivate(priv, context);
+ if (err) {
+ priv->channels.params = old_params;
+ return err;
+ }
+ }
+
+ return 0;
+}
+
static int mlx5e_switch_priv_channels(struct mlx5e_priv *priv,
struct mlx5e_channels *new_chs,
mlx5e_fp_preactivate preactivate,
@@ -2852,16 +2875,12 @@ static int mlx5e_switch_priv_channels(struct mlx5e_priv *priv,
struct net_device *netdev = priv->netdev;
struct mlx5e_channels old_chs;
int carrier_ok;
- bool opened;
int err = 0;
- opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
- if (opened) {
- carrier_ok = netif_carrier_ok(netdev);
- netif_carrier_off(netdev);
+ carrier_ok = netif_carrier_ok(netdev);
+ netif_carrier_off(netdev);
- mlx5e_deactivate_priv_channels(priv);
- }
+ mlx5e_deactivate_priv_channels(priv);
old_chs = priv->channels;
priv->channels = *new_chs;
@@ -2877,19 +2896,15 @@ static int mlx5e_switch_priv_channels(struct mlx5e_priv *priv,
}
}
- if (opened) {
- mlx5e_close_channels(&old_chs);
- priv->profile->update_rx(priv);
- }
+ mlx5e_close_channels(&old_chs);
+ priv->profile->update_rx(priv);
out:
- if (opened) {
- mlx5e_activate_priv_channels(priv);
+ mlx5e_activate_priv_channels(priv);
- /* return carrier back if needed */
- if (carrier_ok)
- netif_carrier_on(netdev);
- }
+ /* return carrier back if needed */
+ if (carrier_ok)
+ netif_carrier_on(netdev);
return err;
}
@@ -2897,27 +2912,19 @@ static int mlx5e_switch_priv_channels(struct mlx5e_priv *priv,
int mlx5e_safe_switch_channels(struct mlx5e_priv *priv,
struct mlx5e_channels *new_chs,
mlx5e_fp_preactivate preactivate,
- void *context)
+ void *context, bool reset)
{
- bool opened;
int err;
- opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
-
- if (opened) {
- err = mlx5e_open_channels(priv, new_chs);
- if (err)
- return err;
- }
+ reset &= test_bit(MLX5E_STATE_OPENED, &priv->state);
+ if (!reset)
+ return mlx5e_switch_priv_params(priv, &new_chs->params, preactivate, context);
+ err = mlx5e_open_channels(priv, new_chs);
+ if (err)
+ return err;
err = mlx5e_switch_priv_channels(priv, new_chs, preactivate, context);
if (err)
- goto err_close;
-
- return 0;
-
-err_close:
- if (opened)
mlx5e_close_channels(new_chs);
return err;
@@ -2928,7 +2935,7 @@ int mlx5e_safe_reopen_channels(struct mlx5e_priv *priv)
struct mlx5e_channels new_channels = {};
new_channels.params = priv->channels.params;
- return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
+ return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, true);
}
void mlx5e_timestamp_init(struct mlx5e_priv *priv)
@@ -3408,7 +3415,7 @@ static int mlx5e_setup_tc_mqprio(struct mlx5e_priv *priv,
new_channels.params.num_tc = tc ? tc : 1;
err = mlx5e_safe_switch_channels(priv, &new_channels,
- mlx5e_num_channels_changed_ctx, NULL);
+ mlx5e_num_channels_changed_ctx, NULL, true);
out:
priv->max_opened_tc = max_t(u8, priv->max_opened_tc,
@@ -3635,7 +3642,7 @@ static int set_feature_lro(struct net_device *netdev, bool enable)
struct mlx5_core_dev *mdev = priv->mdev;
struct mlx5e_channels new_channels = {};
struct mlx5e_params *cur_params;
- bool skip_reset = false;
+ bool reset = true;
int err = 0;
mutex_lock(&priv->state_lock);
@@ -3660,22 +3667,11 @@ static int set_feature_lro(struct net_device *netdev, bool enable)
if (cur_params->rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ) {
if (mlx5e_rx_mpwqe_is_linear_skb(mdev, cur_params, NULL) ==
mlx5e_rx_mpwqe_is_linear_skb(mdev, &new_channels.params, NULL))
- skip_reset = true;
- }
-
- if (skip_reset) {
- struct mlx5e_params old_params;
-
- old_params = *cur_params;
- *cur_params = new_channels.params;
- err = mlx5e_modify_tirs_lro(priv);
- if (err)
- *cur_params = old_params;
- goto out;
+ reset = false;
}
err = mlx5e_safe_switch_channels(priv, &new_channels,
- mlx5e_modify_tirs_lro_ctx, NULL);
+ mlx5e_modify_tirs_lro_ctx, NULL, reset);
out:
mutex_unlock(&priv->state_lock);
return err;
@@ -3906,7 +3902,7 @@ int mlx5e_change_mtu(struct net_device *netdev, int new_mtu,
struct mlx5e_priv *priv = netdev_priv(netdev);
struct mlx5e_channels new_channels = {};
struct mlx5e_params *params;
- bool skip_reset = false;
+ bool reset = true;
int err = 0;
mutex_lock(&priv->state_lock);
@@ -3935,7 +3931,7 @@ int mlx5e_change_mtu(struct net_device *netdev, int new_mtu,
}
if (params->lro_en)
- skip_reset = true;
+ reset = false;
if (params->rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ) {
bool is_linear_old = mlx5e_rx_mpwqe_is_linear_skb(priv->mdev, params, NULL);
@@ -3950,24 +3946,10 @@ int mlx5e_change_mtu(struct net_device *netdev, int new_mtu,
*/
if (!is_linear_old && !is_linear_new && !priv->xsk.refcnt &&
ppw_old == ppw_new)
- skip_reset = true;
- }
-
- if (skip_reset) {
- unsigned int old_mtu = params->sw_mtu;
-
- params->sw_mtu = new_mtu;
- if (preactivate) {
- err = preactivate(priv, NULL);
- if (err) {
- params->sw_mtu = old_mtu;
- goto out;
- }
- }
- goto out;
+ reset = false;
}
- err = mlx5e_safe_switch_channels(priv, &new_channels, preactivate, NULL);
+ err = mlx5e_safe_switch_channels(priv, &new_channels, preactivate, NULL, reset);
out:
netdev->mtu = params->sw_mtu;
@@ -4046,7 +4028,7 @@ int mlx5e_hwstamp_set(struct mlx5e_priv *priv, struct ifreq *ifr)
goto out;
err = mlx5e_safe_switch_channels(priv, &new_channels, mlx5e_ptp_rx_manage_fs_ctx,
- &new_channels.params.ptp_rx);
+ &new_channels.params.ptp_rx, true);
if (err) {
mutex_unlock(&priv->state_lock);
return err;
@@ -4406,9 +4388,10 @@ static void mlx5e_rq_replace_xdp_prog(struct mlx5e_rq *rq, struct bpf_prog *prog
static int mlx5e_xdp_set(struct net_device *netdev, struct bpf_prog *prog)
{
struct mlx5e_priv *priv = netdev_priv(netdev);
+ struct mlx5e_channels new_channels = {};
struct bpf_prog *old_prog;
- bool reset, was_opened;
int err = 0;
+ bool reset;
int i;
mutex_lock(&priv->state_lock);
@@ -4419,43 +4402,29 @@ static int mlx5e_xdp_set(struct net_device *netdev, struct bpf_prog *prog)
goto unlock;
}
- was_opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
/* no need for full reset when exchanging programs */
reset = (!priv->channels.params.xdp_prog || !prog);
- if (was_opened && !reset)
- /* num_channels is invariant here, so we can take the
- * batched reference right upfront.
- */
- bpf_prog_add(prog, priv->channels.num);
-
- if (reset) {
- struct mlx5e_channels new_channels = {};
-
- new_channels.params = priv->channels.params;
- new_channels.params.xdp_prog = prog;
+ new_channels.params = priv->channels.params;
+ new_channels.params.xdp_prog = prog;
+ if (reset)
mlx5e_set_rq_type(priv->mdev, &new_channels.params);
- old_prog = priv->channels.params.xdp_prog;
+ old_prog = priv->channels.params.xdp_prog;
- err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
- if (err)
- goto unlock;
- } else {
- /* exchange programs, extra prog reference we got from caller
- * as long as we don't fail from this point onwards.
- */
- old_prog = xchg(&priv->channels.params.xdp_prog, prog);
- }
+ err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, reset);
+ if (err)
+ goto unlock;
if (old_prog)
bpf_prog_put(old_prog);
- if (!was_opened || reset)
+ if (!test_bit(MLX5E_STATE_OPENED, &priv->state) || reset)
goto unlock;
/* exchanging programs w/o reset, we update ref counts on behalf
* of the channels RQs here.
*/
+ bpf_prog_add(prog, priv->channels.num);
for (i = 0; i < priv->channels.num; i++) {
struct mlx5e_channel *c = priv->channels.c[i];
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
index 9555127ce7e7..df409111ef3a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
@@ -489,7 +489,7 @@ static int mlx5i_change_mtu(struct net_device *netdev, int new_mtu)
new_channels.params = priv->channels.params;
new_channels.params.sw_mtu = new_mtu;
- err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL);
+ err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, true);
if (err)
goto out;
--
2.30.2
^ permalink raw reply related [flat|nested] 16+ messages in thread* [net-next 09/14] net/mlx5e: Cleanup safe switch channels API by passing params
2021-04-16 18:54 [pull request][net-next 00/14] mlx5 updates 2021-04-16 Saeed Mahameed
` (7 preceding siblings ...)
2021-04-16 18:54 ` [net-next 08/14] net/mlx5e: Refactor on-the-fly configuration changes Saeed Mahameed
@ 2021-04-16 18:54 ` Saeed Mahameed
2021-04-16 18:54 ` [net-next 10/14] net/mlx5: Allocate FC bulk structs with kvzalloc() instead of kzalloc() Saeed Mahameed
` (4 subsequent siblings)
13 siblings, 0 replies; 16+ messages in thread
From: Saeed Mahameed @ 2021-04-16 18:54 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: Tariq Toukan, netdev, Maxim Mikityanskiy, Saeed Mahameed
From: Maxim Mikityanskiy <maximmi@nvidia.com>
mlx5e_safe_switch_channels accepts new_chs as a parameter and opens new
channels in place, then copying them to priv->channels. It requires all
the callers to allocate space for this temporary storage of the new
channels.
This commit cleans up the API by replacing new_chs with new_params, a
meaningful subset of new_chs to be filled by the caller. The temporary
space for the new channels is allocated inside mlx5e_safe_switch_params
(a new name for mlx5e_safe_switch_channels). An extra copy of params is
made, but since it's control flow, it's not critical.
Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en.h | 8 +-
.../ethernet/mellanox/mlx5/core/en_dcbnl.c | 15 ++-
.../ethernet/mellanox/mlx5/core/en_ethtool.c | 98 +++++++++---------
.../net/ethernet/mellanox/mlx5/core/en_main.c | 99 +++++++++----------
.../ethernet/mellanox/mlx5/core/ipoib/ipoib.c | 10 +-
5 files changed, 114 insertions(+), 116 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index d9ed20a4db53..b636d63358d2 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -1018,10 +1018,10 @@ int fn##_ctx(struct mlx5e_priv *priv, void *context) \
return fn(priv); \
}
int mlx5e_safe_reopen_channels(struct mlx5e_priv *priv);
-int mlx5e_safe_switch_channels(struct mlx5e_priv *priv,
- struct mlx5e_channels *new_chs,
- mlx5e_fp_preactivate preactivate,
- void *context, bool reset);
+int mlx5e_safe_switch_params(struct mlx5e_priv *priv,
+ struct mlx5e_params *new_params,
+ mlx5e_fp_preactivate preactivate,
+ void *context, bool reset);
int mlx5e_update_tx_netdev_queues(struct mlx5e_priv *priv);
int mlx5e_num_channels_changed(struct mlx5e_priv *priv);
int mlx5e_num_channels_changed_ctx(struct mlx5e_priv *priv, void *context);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
index 0b0273fac6e3..a4c8d8d00d5a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
@@ -1149,24 +1149,23 @@ static int mlx5e_update_trust_state_hw(struct mlx5e_priv *priv, void *context)
static int mlx5e_set_trust_state(struct mlx5e_priv *priv, u8 trust_state)
{
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
bool reset = true;
int err;
mutex_lock(&priv->state_lock);
- new_channels.params = priv->channels.params;
- mlx5e_params_calc_trust_tx_min_inline_mode(priv->mdev, &new_channels.params,
+ new_params = priv->channels.params;
+ mlx5e_params_calc_trust_tx_min_inline_mode(priv->mdev, &new_params,
trust_state);
/* Skip if tx_min_inline is the same */
- if (new_channels.params.tx_min_inline_mode ==
- priv->channels.params.tx_min_inline_mode)
+ if (new_params.tx_min_inline_mode == priv->channels.params.tx_min_inline_mode)
reset = false;
- err = mlx5e_safe_switch_channels(priv, &new_channels,
- mlx5e_update_trust_state_hw,
- &trust_state, reset);
+ err = mlx5e_safe_switch_params(priv, &new_params,
+ mlx5e_update_trust_state_hw,
+ &trust_state, reset);
mutex_unlock(&priv->state_lock);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
index 82994175aded..ef4a330c4cfa 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -326,7 +326,7 @@ static void mlx5e_get_ringparam(struct net_device *dev,
int mlx5e_ethtool_set_ringparam(struct mlx5e_priv *priv,
struct ethtool_ringparam *param)
{
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
u8 log_rq_size;
u8 log_sq_size;
int err = 0;
@@ -365,15 +365,15 @@ int mlx5e_ethtool_set_ringparam(struct mlx5e_priv *priv,
mutex_lock(&priv->state_lock);
- new_channels.params = priv->channels.params;
- new_channels.params.log_rq_mtu_frames = log_rq_size;
- new_channels.params.log_sq_size = log_sq_size;
+ new_params = priv->channels.params;
+ new_params.log_rq_mtu_frames = log_rq_size;
+ new_params.log_sq_size = log_sq_size;
- err = mlx5e_validate_params(priv->mdev, &new_channels.params);
+ err = mlx5e_validate_params(priv->mdev, &new_params);
if (err)
goto unlock;
- err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, true);
+ err = mlx5e_safe_switch_params(priv, &new_params, NULL, NULL, true);
unlock:
mutex_unlock(&priv->state_lock);
@@ -418,7 +418,7 @@ int mlx5e_ethtool_set_channels(struct mlx5e_priv *priv,
{
struct mlx5e_params *cur_params = &priv->channels.params;
unsigned int count = ch->combined_count;
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
bool arfs_enabled;
bool opened;
int err = 0;
@@ -455,8 +455,8 @@ int mlx5e_ethtool_set_channels(struct mlx5e_priv *priv,
goto out;
}
- new_channels.params = *cur_params;
- new_channels.params.num_channels = count;
+ new_params = *cur_params;
+ new_params.num_channels = count;
opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
@@ -465,8 +465,8 @@ int mlx5e_ethtool_set_channels(struct mlx5e_priv *priv,
mlx5e_arfs_disable(priv);
/* Switch to new channels, set new parameters and close old ones */
- err = mlx5e_safe_switch_channels(priv, &new_channels,
- mlx5e_num_channels_changed_ctx, NULL, true);
+ err = mlx5e_safe_switch_params(priv, &new_params,
+ mlx5e_num_channels_changed_ctx, NULL, true);
if (arfs_enabled) {
int err2 = mlx5e_arfs_enable(priv);
@@ -561,7 +561,7 @@ int mlx5e_ethtool_set_coalesce(struct mlx5e_priv *priv,
{
struct dim_cq_moder *rx_moder, *tx_moder;
struct mlx5_core_dev *mdev = priv->mdev;
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
bool reset_rx, reset_tx;
bool reset = true;
int err = 0;
@@ -584,32 +584,32 @@ int mlx5e_ethtool_set_coalesce(struct mlx5e_priv *priv,
}
mutex_lock(&priv->state_lock);
- new_channels.params = priv->channels.params;
+ new_params = priv->channels.params;
- rx_moder = &new_channels.params.rx_cq_moderation;
+ rx_moder = &new_params.rx_cq_moderation;
rx_moder->usec = coal->rx_coalesce_usecs;
rx_moder->pkts = coal->rx_max_coalesced_frames;
- new_channels.params.rx_dim_enabled = !!coal->use_adaptive_rx_coalesce;
+ new_params.rx_dim_enabled = !!coal->use_adaptive_rx_coalesce;
- tx_moder = &new_channels.params.tx_cq_moderation;
+ tx_moder = &new_params.tx_cq_moderation;
tx_moder->usec = coal->tx_coalesce_usecs;
tx_moder->pkts = coal->tx_max_coalesced_frames;
- new_channels.params.tx_dim_enabled = !!coal->use_adaptive_tx_coalesce;
+ new_params.tx_dim_enabled = !!coal->use_adaptive_tx_coalesce;
reset_rx = !!coal->use_adaptive_rx_coalesce != priv->channels.params.rx_dim_enabled;
reset_tx = !!coal->use_adaptive_tx_coalesce != priv->channels.params.tx_dim_enabled;
if (reset_rx) {
- u8 mode = MLX5E_GET_PFLAG(&new_channels.params,
+ u8 mode = MLX5E_GET_PFLAG(&new_params,
MLX5E_PFLAG_RX_CQE_BASED_MODER);
- mlx5e_reset_rx_moderation(&new_channels.params, mode);
+ mlx5e_reset_rx_moderation(&new_params, mode);
}
if (reset_tx) {
- u8 mode = MLX5E_GET_PFLAG(&new_channels.params,
+ u8 mode = MLX5E_GET_PFLAG(&new_params,
MLX5E_PFLAG_TX_CQE_BASED_MODER);
- mlx5e_reset_tx_moderation(&new_channels.params, mode);
+ mlx5e_reset_tx_moderation(&new_params, mode);
}
/* If DIM state hasn't changed, it's possible to modify interrupt
@@ -623,7 +623,7 @@ int mlx5e_ethtool_set_coalesce(struct mlx5e_priv *priv,
reset = false;
}
- err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, reset);
+ err = mlx5e_safe_switch_params(priv, &new_params, NULL, NULL, reset);
mutex_unlock(&priv->state_lock);
return err;
@@ -1843,7 +1843,7 @@ static int set_pflag_cqe_based_moder(struct net_device *netdev, bool enable,
{
struct mlx5e_priv *priv = netdev_priv(netdev);
struct mlx5_core_dev *mdev = priv->mdev;
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
bool mode_changed;
u8 cq_period_mode, current_cq_period_mode;
@@ -1862,13 +1862,13 @@ static int set_pflag_cqe_based_moder(struct net_device *netdev, bool enable,
if (!mode_changed)
return 0;
- new_channels.params = priv->channels.params;
+ new_params = priv->channels.params;
if (is_rx_cq)
- mlx5e_set_rx_cq_mode_params(&new_channels.params, cq_period_mode);
+ mlx5e_set_rx_cq_mode_params(&new_params, cq_period_mode);
else
- mlx5e_set_tx_cq_mode_params(&new_channels.params, cq_period_mode);
+ mlx5e_set_tx_cq_mode_params(&new_params, cq_period_mode);
- return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, true);
+ return mlx5e_safe_switch_params(priv, &new_params, NULL, NULL, true);
}
static int set_pflag_tx_cqe_based_moder(struct net_device *netdev, bool enable)
@@ -1884,7 +1884,7 @@ static int set_pflag_rx_cqe_based_moder(struct net_device *netdev, bool enable)
int mlx5e_modify_rx_cqe_compression_locked(struct mlx5e_priv *priv, bool new_val)
{
bool curr_val = MLX5E_GET_PFLAG(&priv->channels.params, MLX5E_PFLAG_RX_CQE_COMPRESS);
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
int err = 0;
if (!MLX5_CAP_GEN(priv->mdev, cqe_compression))
@@ -1893,16 +1893,16 @@ int mlx5e_modify_rx_cqe_compression_locked(struct mlx5e_priv *priv, bool new_val
if (curr_val == new_val)
return 0;
- new_channels.params = priv->channels.params;
- MLX5E_SET_PFLAG(&new_channels.params, MLX5E_PFLAG_RX_CQE_COMPRESS, new_val);
+ new_params = priv->channels.params;
+ MLX5E_SET_PFLAG(&new_params, MLX5E_PFLAG_RX_CQE_COMPRESS, new_val);
if (priv->tstamp.rx_filter != HWTSTAMP_FILTER_NONE)
- new_channels.params.ptp_rx = new_val;
+ new_params.ptp_rx = new_val;
- if (new_channels.params.ptp_rx == priv->channels.params.ptp_rx)
- err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, true);
+ if (new_params.ptp_rx == priv->channels.params.ptp_rx)
+ err = mlx5e_safe_switch_params(priv, &new_params, NULL, NULL, true);
else
- err = mlx5e_safe_switch_channels(priv, &new_channels, mlx5e_ptp_rx_manage_fs_ctx,
- &new_channels.params.ptp_rx, true);
+ err = mlx5e_safe_switch_params(priv, &new_params, mlx5e_ptp_rx_manage_fs_ctx,
+ &new_params.ptp_rx, true);
if (err)
return err;
@@ -1936,7 +1936,7 @@ static int set_pflag_rx_striding_rq(struct net_device *netdev, bool enable)
{
struct mlx5e_priv *priv = netdev_priv(netdev);
struct mlx5_core_dev *mdev = priv->mdev;
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
if (enable) {
if (!mlx5e_check_fragmented_striding_rq_cap(mdev))
@@ -1948,12 +1948,12 @@ static int set_pflag_rx_striding_rq(struct net_device *netdev, bool enable)
return -EINVAL;
}
- new_channels.params = priv->channels.params;
+ new_params = priv->channels.params;
- MLX5E_SET_PFLAG(&new_channels.params, MLX5E_PFLAG_RX_STRIDING_RQ, enable);
- mlx5e_set_rq_type(mdev, &new_channels.params);
+ MLX5E_SET_PFLAG(&new_params, MLX5E_PFLAG_RX_STRIDING_RQ, enable);
+ mlx5e_set_rq_type(mdev, &new_params);
- return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, true);
+ return mlx5e_safe_switch_params(priv, &new_params, NULL, NULL, true);
}
static int set_pflag_rx_no_csum_complete(struct net_device *netdev, bool enable)
@@ -1982,16 +1982,16 @@ static int set_pflag_tx_mpwqe_common(struct net_device *netdev, u32 flag, bool e
{
struct mlx5e_priv *priv = netdev_priv(netdev);
struct mlx5_core_dev *mdev = priv->mdev;
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
if (enable && !MLX5_CAP_ETH(mdev, enhanced_multi_pkt_send_wqe))
return -EOPNOTSUPP;
- new_channels.params = priv->channels.params;
+ new_params = priv->channels.params;
- MLX5E_SET_PFLAG(&new_channels.params, flag, enable);
+ MLX5E_SET_PFLAG(&new_params, flag, enable);
- return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, true);
+ return mlx5e_safe_switch_params(priv, &new_params, NULL, NULL, true);
}
static int set_pflag_xdp_tx_mpwqe(struct net_device *netdev, bool enable)
@@ -2008,7 +2008,7 @@ static int set_pflag_tx_port_ts(struct net_device *netdev, bool enable)
{
struct mlx5e_priv *priv = netdev_priv(netdev);
struct mlx5_core_dev *mdev = priv->mdev;
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
int err;
if (!MLX5_CAP_GEN(mdev, ts_cqe_to_dest_cqn))
@@ -2024,15 +2024,15 @@ static int set_pflag_tx_port_ts(struct net_device *netdev, bool enable)
return -EINVAL;
}
- new_channels.params = priv->channels.params;
- MLX5E_SET_PFLAG(&new_channels.params, MLX5E_PFLAG_TX_PORT_TS, enable);
+ new_params = priv->channels.params;
+ MLX5E_SET_PFLAG(&new_params, MLX5E_PFLAG_TX_PORT_TS, enable);
/* No need to verify SQ stop room as
* ptpsq.txqsq.stop_room <= generic_sq->stop_room, and both
* has the same log_sq_size.
*/
- err = mlx5e_safe_switch_channels(priv, &new_channels,
- mlx5e_num_channels_changed_ctx, NULL, true);
+ err = mlx5e_safe_switch_params(priv, &new_params,
+ mlx5e_num_channels_changed_ctx, NULL, true);
if (!err)
priv->tx_ptp_opened = true;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 7686d4997696..feb347e81448 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -2909,33 +2909,32 @@ static int mlx5e_switch_priv_channels(struct mlx5e_priv *priv,
return err;
}
-int mlx5e_safe_switch_channels(struct mlx5e_priv *priv,
- struct mlx5e_channels *new_chs,
- mlx5e_fp_preactivate preactivate,
- void *context, bool reset)
+int mlx5e_safe_switch_params(struct mlx5e_priv *priv,
+ struct mlx5e_params *params,
+ mlx5e_fp_preactivate preactivate,
+ void *context, bool reset)
{
+ struct mlx5e_channels new_chs = {};
int err;
reset &= test_bit(MLX5E_STATE_OPENED, &priv->state);
if (!reset)
- return mlx5e_switch_priv_params(priv, &new_chs->params, preactivate, context);
+ return mlx5e_switch_priv_params(priv, params, preactivate, context);
- err = mlx5e_open_channels(priv, new_chs);
+ new_chs.params = *params;
+ err = mlx5e_open_channels(priv, &new_chs);
if (err)
return err;
- err = mlx5e_switch_priv_channels(priv, new_chs, preactivate, context);
+ err = mlx5e_switch_priv_channels(priv, &new_chs, preactivate, context);
if (err)
- mlx5e_close_channels(new_chs);
+ mlx5e_close_channels(&new_chs);
return err;
}
int mlx5e_safe_reopen_channels(struct mlx5e_priv *priv)
{
- struct mlx5e_channels new_channels = {};
-
- new_channels.params = priv->channels.params;
- return mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, true);
+ return mlx5e_safe_switch_params(priv, &priv->channels.params, NULL, NULL, true);
}
void mlx5e_timestamp_init(struct mlx5e_priv *priv)
@@ -3392,7 +3391,7 @@ static int mlx5e_modify_channels_vsd(struct mlx5e_channels *chs, bool vsd)
static int mlx5e_setup_tc_mqprio(struct mlx5e_priv *priv,
struct tc_mqprio_qopt *mqprio)
{
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
u8 tc = mqprio->num_tc;
int err = 0;
@@ -3411,11 +3410,11 @@ static int mlx5e_setup_tc_mqprio(struct mlx5e_priv *priv,
goto out;
}
- new_channels.params = priv->channels.params;
- new_channels.params.num_tc = tc ? tc : 1;
+ new_params = priv->channels.params;
+ new_params.num_tc = tc ? tc : 1;
- err = mlx5e_safe_switch_channels(priv, &new_channels,
- mlx5e_num_channels_changed_ctx, NULL, true);
+ err = mlx5e_safe_switch_params(priv, &new_params,
+ mlx5e_num_channels_changed_ctx, NULL, true);
out:
priv->max_opened_tc = max_t(u8, priv->max_opened_tc,
@@ -3640,8 +3639,8 @@ static int set_feature_lro(struct net_device *netdev, bool enable)
{
struct mlx5e_priv *priv = netdev_priv(netdev);
struct mlx5_core_dev *mdev = priv->mdev;
- struct mlx5e_channels new_channels = {};
struct mlx5e_params *cur_params;
+ struct mlx5e_params new_params;
bool reset = true;
int err = 0;
@@ -3661,17 +3660,17 @@ static int set_feature_lro(struct net_device *netdev, bool enable)
goto out;
}
- new_channels.params = *cur_params;
- new_channels.params.lro_en = enable;
+ new_params = *cur_params;
+ new_params.lro_en = enable;
if (cur_params->rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ) {
if (mlx5e_rx_mpwqe_is_linear_skb(mdev, cur_params, NULL) ==
- mlx5e_rx_mpwqe_is_linear_skb(mdev, &new_channels.params, NULL))
+ mlx5e_rx_mpwqe_is_linear_skb(mdev, &new_params, NULL))
reset = false;
}
- err = mlx5e_safe_switch_channels(priv, &new_channels,
- mlx5e_modify_tirs_lro_ctx, NULL, reset);
+ err = mlx5e_safe_switch_params(priv, &new_params,
+ mlx5e_modify_tirs_lro_ctx, NULL, reset);
out:
mutex_unlock(&priv->state_lock);
return err;
@@ -3900,7 +3899,7 @@ int mlx5e_change_mtu(struct net_device *netdev, int new_mtu,
mlx5e_fp_preactivate preactivate)
{
struct mlx5e_priv *priv = netdev_priv(netdev);
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
struct mlx5e_params *params;
bool reset = true;
int err = 0;
@@ -3909,14 +3908,14 @@ int mlx5e_change_mtu(struct net_device *netdev, int new_mtu,
params = &priv->channels.params;
- new_channels.params = *params;
- new_channels.params.sw_mtu = new_mtu;
- err = mlx5e_validate_params(priv->mdev, &new_channels.params);
+ new_params = *params;
+ new_params.sw_mtu = new_mtu;
+ err = mlx5e_validate_params(priv->mdev, &new_params);
if (err)
goto out;
if (params->xdp_prog &&
- !mlx5e_rx_is_linear_skb(&new_channels.params, NULL)) {
+ !mlx5e_rx_is_linear_skb(&new_params, NULL)) {
netdev_err(netdev, "MTU(%d) > %d is not allowed while XDP enabled\n",
new_mtu, mlx5e_xdp_max_mtu(params, NULL));
err = -EINVAL;
@@ -3925,7 +3924,7 @@ int mlx5e_change_mtu(struct net_device *netdev, int new_mtu,
if (priv->xsk.refcnt &&
!mlx5e_xsk_validate_mtu(netdev, &priv->channels,
- &new_channels.params, priv->mdev)) {
+ &new_params, priv->mdev)) {
err = -EINVAL;
goto out;
}
@@ -3936,9 +3935,9 @@ int mlx5e_change_mtu(struct net_device *netdev, int new_mtu,
if (params->rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ) {
bool is_linear_old = mlx5e_rx_mpwqe_is_linear_skb(priv->mdev, params, NULL);
bool is_linear_new = mlx5e_rx_mpwqe_is_linear_skb(priv->mdev,
- &new_channels.params, NULL);
+ &new_params, NULL);
u8 ppw_old = mlx5e_mpwqe_log_pkts_per_wqe(params, NULL);
- u8 ppw_new = mlx5e_mpwqe_log_pkts_per_wqe(&new_channels.params, NULL);
+ u8 ppw_new = mlx5e_mpwqe_log_pkts_per_wqe(&new_params, NULL);
/* Always reset in linear mode - hw_mtu is used in data path.
* Check that the mode was non-linear and didn't change.
@@ -3949,7 +3948,7 @@ int mlx5e_change_mtu(struct net_device *netdev, int new_mtu,
reset = false;
}
- err = mlx5e_safe_switch_channels(priv, &new_channels, preactivate, NULL, reset);
+ err = mlx5e_safe_switch_params(priv, &new_params, preactivate, NULL, reset);
out:
netdev->mtu = params->sw_mtu;
@@ -3971,7 +3970,7 @@ int mlx5e_ptp_rx_manage_fs_ctx(struct mlx5e_priv *priv, void *ctx)
int mlx5e_hwstamp_set(struct mlx5e_priv *priv, struct ifreq *ifr)
{
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
struct hwtstamp_config config;
bool rx_cqe_compress_def;
int err;
@@ -3993,13 +3992,13 @@ int mlx5e_hwstamp_set(struct mlx5e_priv *priv, struct ifreq *ifr)
}
mutex_lock(&priv->state_lock);
- new_channels.params = priv->channels.params;
+ new_params = priv->channels.params;
rx_cqe_compress_def = priv->channels.params.rx_cqe_compress_def;
/* RX HW timestamp */
switch (config.rx_filter) {
case HWTSTAMP_FILTER_NONE:
- new_channels.params.ptp_rx = false;
+ new_params.ptp_rx = false;
break;
case HWTSTAMP_FILTER_ALL:
case HWTSTAMP_FILTER_SOME:
@@ -4016,7 +4015,7 @@ int mlx5e_hwstamp_set(struct mlx5e_priv *priv, struct ifreq *ifr)
case HWTSTAMP_FILTER_PTP_V2_SYNC:
case HWTSTAMP_FILTER_PTP_V2_DELAY_REQ:
case HWTSTAMP_FILTER_NTP_ALL:
- new_channels.params.ptp_rx = rx_cqe_compress_def;
+ new_params.ptp_rx = rx_cqe_compress_def;
config.rx_filter = HWTSTAMP_FILTER_ALL;
break;
default:
@@ -4024,11 +4023,11 @@ int mlx5e_hwstamp_set(struct mlx5e_priv *priv, struct ifreq *ifr)
return -ERANGE;
}
- if (new_channels.params.ptp_rx == priv->channels.params.ptp_rx)
+ if (new_params.ptp_rx == priv->channels.params.ptp_rx)
goto out;
- err = mlx5e_safe_switch_channels(priv, &new_channels, mlx5e_ptp_rx_manage_fs_ctx,
- &new_channels.params.ptp_rx, true);
+ err = mlx5e_safe_switch_params(priv, &new_params, mlx5e_ptp_rx_manage_fs_ctx,
+ &new_params.ptp_rx, true);
if (err) {
mutex_unlock(&priv->state_lock);
return err;
@@ -4346,7 +4345,7 @@ static void mlx5e_tx_timeout(struct net_device *dev, unsigned int txqueue)
static int mlx5e_xdp_allowed(struct mlx5e_priv *priv, struct bpf_prog *prog)
{
struct net_device *netdev = priv->netdev;
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
if (priv->channels.params.lro_en) {
netdev_warn(netdev, "can't set XDP while LRO is on, disable LRO first\n");
@@ -4359,16 +4358,16 @@ static int mlx5e_xdp_allowed(struct mlx5e_priv *priv, struct bpf_prog *prog)
return -EINVAL;
}
- new_channels.params = priv->channels.params;
- new_channels.params.xdp_prog = prog;
+ new_params = priv->channels.params;
+ new_params.xdp_prog = prog;
/* No XSK params: AF_XDP can't be enabled yet at the point of setting
* the XDP program.
*/
- if (!mlx5e_rx_is_linear_skb(&new_channels.params, NULL)) {
+ if (!mlx5e_rx_is_linear_skb(&new_params, NULL)) {
netdev_warn(netdev, "XDP is not allowed with MTU(%d) > %d\n",
- new_channels.params.sw_mtu,
- mlx5e_xdp_max_mtu(&new_channels.params, NULL));
+ new_params.sw_mtu,
+ mlx5e_xdp_max_mtu(&new_params, NULL));
return -EINVAL;
}
@@ -4388,7 +4387,7 @@ static void mlx5e_rq_replace_xdp_prog(struct mlx5e_rq *rq, struct bpf_prog *prog
static int mlx5e_xdp_set(struct net_device *netdev, struct bpf_prog *prog)
{
struct mlx5e_priv *priv = netdev_priv(netdev);
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
struct bpf_prog *old_prog;
int err = 0;
bool reset;
@@ -4405,13 +4404,13 @@ static int mlx5e_xdp_set(struct net_device *netdev, struct bpf_prog *prog)
/* no need for full reset when exchanging programs */
reset = (!priv->channels.params.xdp_prog || !prog);
- new_channels.params = priv->channels.params;
- new_channels.params.xdp_prog = prog;
+ new_params = priv->channels.params;
+ new_params.xdp_prog = prog;
if (reset)
- mlx5e_set_rq_type(priv->mdev, &new_channels.params);
+ mlx5e_set_rq_type(priv->mdev, &new_params);
old_prog = priv->channels.params.xdp_prog;
- err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, reset);
+ err = mlx5e_safe_switch_params(priv, &new_params, NULL, NULL, reset);
if (err)
goto unlock;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
index df409111ef3a..612a7f69366d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
@@ -481,19 +481,19 @@ static const struct mlx5e_profile mlx5i_nic_profile = {
static int mlx5i_change_mtu(struct net_device *netdev, int new_mtu)
{
struct mlx5e_priv *priv = mlx5i_epriv(netdev);
- struct mlx5e_channels new_channels = {};
+ struct mlx5e_params new_params;
int err = 0;
mutex_lock(&priv->state_lock);
- new_channels.params = priv->channels.params;
- new_channels.params.sw_mtu = new_mtu;
+ new_params = priv->channels.params;
+ new_params.sw_mtu = new_mtu;
- err = mlx5e_safe_switch_channels(priv, &new_channels, NULL, NULL, true);
+ err = mlx5e_safe_switch_params(priv, &new_params, NULL, NULL, true);
if (err)
goto out;
- netdev->mtu = new_channels.params.sw_mtu;
+ netdev->mtu = new_params.sw_mtu;
out:
mutex_unlock(&priv->state_lock);
--
2.30.2
^ permalink raw reply related [flat|nested] 16+ messages in thread* [net-next 10/14] net/mlx5: Allocate FC bulk structs with kvzalloc() instead of kzalloc()
2021-04-16 18:54 [pull request][net-next 00/14] mlx5 updates 2021-04-16 Saeed Mahameed
` (8 preceding siblings ...)
2021-04-16 18:54 ` [net-next 09/14] net/mlx5e: Cleanup safe switch channels API by passing params Saeed Mahameed
@ 2021-04-16 18:54 ` Saeed Mahameed
2021-04-16 18:54 ` [net-next 11/14] net/mlx5: Add register layout to support extended link state Saeed Mahameed
` (3 subsequent siblings)
13 siblings, 0 replies; 16+ messages in thread
From: Saeed Mahameed @ 2021-04-16 18:54 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: Tariq Toukan, netdev, Maor Dickman, Saeed Mahameed
From: Maor Dickman <maord@nvidia.com>
The bulk size is larger than 16K so use kvzalloc().
The bulk bitmask upper size limit is 16K so use kvcalloc().
Signed-off-by: Maor Dickman <maord@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../ethernet/mellanox/mlx5/core/fs_counters.c | 16 ++++++++--------
1 file changed, 8 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c
index f43caefd07a1..18e5aec14641 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c
@@ -497,13 +497,13 @@ static struct mlx5_fc_bulk *mlx5_fc_bulk_create(struct mlx5_core_dev *dev)
alloc_bitmask = MLX5_CAP_GEN(dev, flow_counter_bulk_alloc);
bulk_len = alloc_bitmask > 0 ? MLX5_FC_BULK_NUM_FCS(alloc_bitmask) : 1;
- bulk = kzalloc(sizeof(*bulk) + bulk_len * sizeof(struct mlx5_fc),
- GFP_KERNEL);
+ bulk = kvzalloc(sizeof(*bulk) + bulk_len * sizeof(struct mlx5_fc),
+ GFP_KERNEL);
if (!bulk)
goto err_alloc_bulk;
- bulk->bitmask = kcalloc(BITS_TO_LONGS(bulk_len), sizeof(unsigned long),
- GFP_KERNEL);
+ bulk->bitmask = kvcalloc(BITS_TO_LONGS(bulk_len), sizeof(unsigned long),
+ GFP_KERNEL);
if (!bulk->bitmask)
goto err_alloc_bitmask;
@@ -521,9 +521,9 @@ static struct mlx5_fc_bulk *mlx5_fc_bulk_create(struct mlx5_core_dev *dev)
return bulk;
err_mlx5_cmd_bulk_alloc:
- kfree(bulk->bitmask);
+ kvfree(bulk->bitmask);
err_alloc_bitmask:
- kfree(bulk);
+ kvfree(bulk);
err_alloc_bulk:
return ERR_PTR(err);
}
@@ -537,8 +537,8 @@ mlx5_fc_bulk_destroy(struct mlx5_core_dev *dev, struct mlx5_fc_bulk *bulk)
}
mlx5_cmd_fc_free(dev, bulk->base_id);
- kfree(bulk->bitmask);
- kfree(bulk);
+ kvfree(bulk->bitmask);
+ kvfree(bulk);
return 0;
}
--
2.30.2
^ permalink raw reply related [flat|nested] 16+ messages in thread* [net-next 11/14] net/mlx5: Add register layout to support extended link state
2021-04-16 18:54 [pull request][net-next 00/14] mlx5 updates 2021-04-16 Saeed Mahameed
` (9 preceding siblings ...)
2021-04-16 18:54 ` [net-next 10/14] net/mlx5: Allocate FC bulk structs with kvzalloc() instead of kzalloc() Saeed Mahameed
@ 2021-04-16 18:54 ` Saeed Mahameed
2021-04-16 18:54 ` [net-next 12/14] net/mlx5e: Add ethtool " Saeed Mahameed
` (2 subsequent siblings)
13 siblings, 0 replies; 16+ messages in thread
From: Saeed Mahameed @ 2021-04-16 18:54 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: Tariq Toukan, netdev, Moshe Tal, Saeed Mahameed
From: Moshe Tal <moshet@nvidia.com>
Add needed structure layouts and defines for pddr register
(Port Diagnostics Database Register) and the troublshooting page.
This will be used to get extended link state from the monitor opcode
bits.
Signed-off-by: Moshe Tal <moshet@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
include/linux/mlx5/driver.h | 1 +
include/linux/mlx5/mlx5_ifc.h | 50 +++++++++++++++++++++++++++++++++++
2 files changed, 51 insertions(+)
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index 2da953ad02ed..4e531c2aab52 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -127,6 +127,7 @@ enum {
MLX5_REG_PELC = 0x500e,
MLX5_REG_PVLC = 0x500f,
MLX5_REG_PCMR = 0x5041,
+ MLX5_REG_PDDR = 0x5031,
MLX5_REG_PMLP = 0x5002,
MLX5_REG_PPLM = 0x5023,
MLX5_REG_PCAM = 0x507f,
diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h
index 1599deee0456..f2c51d6833c6 100644
--- a/include/linux/mlx5/mlx5_ifc.h
+++ b/include/linux/mlx5/mlx5_ifc.h
@@ -9956,6 +9956,53 @@ struct mlx5_ifc_mirc_reg_bits {
u8 reserved_at_20[0x20];
};
+struct mlx5_ifc_pddr_monitor_opcode_bits {
+ u8 reserved_at_0[0x10];
+ u8 monitor_opcode[0x10];
+};
+
+union mlx5_ifc_pddr_troubleshooting_page_status_opcode_auto_bits {
+ struct mlx5_ifc_pddr_monitor_opcode_bits pddr_monitor_opcode;
+ u8 reserved_at_0[0x20];
+};
+
+enum {
+ /* Monitor opcodes */
+ MLX5_PDDR_REG_TRBLSH_GROUP_OPCODE_MONITOR = 0x0,
+};
+
+struct mlx5_ifc_pddr_troubleshooting_page_bits {
+ u8 reserved_at_0[0x10];
+ u8 group_opcode[0x10];
+
+ union mlx5_ifc_pddr_troubleshooting_page_status_opcode_auto_bits status_opcode;
+
+ u8 reserved_at_40[0x20];
+
+ u8 status_message[59][0x20];
+};
+
+union mlx5_ifc_pddr_reg_page_data_auto_bits {
+ struct mlx5_ifc_pddr_troubleshooting_page_bits pddr_troubleshooting_page;
+ u8 reserved_at_0[0x7c0];
+};
+
+enum {
+ MLX5_PDDR_REG_PAGE_SELECT_TROUBLESHOOTING_INFO_PAGE = 0x1,
+};
+
+struct mlx5_ifc_pddr_reg_bits {
+ u8 reserved_at_0[0x8];
+ u8 local_port[0x8];
+ u8 pnat[0x2];
+ u8 reserved_at_12[0xe];
+
+ u8 reserved_at_20[0x18];
+ u8 page_select[0x8];
+
+ union mlx5_ifc_pddr_reg_page_data_auto_bits page_data;
+};
+
union mlx5_ifc_ports_control_registers_document_bits {
struct mlx5_ifc_bufferx_reg_bits bufferx_reg;
struct mlx5_ifc_eth_2819_cntrs_grp_data_layout_bits eth_2819_cntrs_grp_data_layout;
@@ -9970,6 +10017,9 @@ union mlx5_ifc_ports_control_registers_document_bits {
struct mlx5_ifc_pamp_reg_bits pamp_reg;
struct mlx5_ifc_paos_reg_bits paos_reg;
struct mlx5_ifc_pcap_reg_bits pcap_reg;
+ struct mlx5_ifc_pddr_monitor_opcode_bits pddr_monitor_opcode;
+ struct mlx5_ifc_pddr_reg_bits pddr_reg;
+ struct mlx5_ifc_pddr_troubleshooting_page_bits pddr_troubleshooting_page;
struct mlx5_ifc_peir_reg_bits peir_reg;
struct mlx5_ifc_pelc_reg_bits pelc_reg;
struct mlx5_ifc_pfcc_reg_bits pfcc_reg;
--
2.30.2
^ permalink raw reply related [flat|nested] 16+ messages in thread* [net-next 12/14] net/mlx5e: Add ethtool extended link state
2021-04-16 18:54 [pull request][net-next 00/14] mlx5 updates 2021-04-16 Saeed Mahameed
` (10 preceding siblings ...)
2021-04-16 18:54 ` [net-next 11/14] net/mlx5: Add register layout to support extended link state Saeed Mahameed
@ 2021-04-16 18:54 ` Saeed Mahameed
2021-04-16 18:54 ` [net-next 13/14] net/mlx5: Add helper to initialize 1PPS Saeed Mahameed
2021-04-16 18:54 ` [net-next 14/14] net/mlx5: Enhance diagnostics info for TX/RX reporters Saeed Mahameed
13 siblings, 0 replies; 16+ messages in thread
From: Saeed Mahameed @ 2021-04-16 18:54 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: Tariq Toukan, netdev, Moshe Tal, Saeed Mahameed
From: Moshe Tal <moshet@nvidia.com>
In case the interface was set up but cannot establish the link, ethtool
will print more information to help the user troubleshoot the state.
For example, no link due to missing cable:
$ ethtool eth1
...
Link detected: no (No cable)
Beside the general extended state, drivers can pass additional
information about the link state using the sub-state field. For example:
$ ethtool eth1
...
Link detected: no (Autoneg, No partner detected)
The extended state is available only for specific cases, in other cases
ethtool with print only "Link detected: no" as before
Signed-off-by: Moshe Tal <moshet@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../ethernet/mellanox/mlx5/core/en_ethtool.c | 171 ++++++++++++++++++
1 file changed, 171 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
index ef4a330c4cfa..e1d3cd1f1f11 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -2125,12 +2125,183 @@ int mlx5e_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd)
return mlx5e_ethtool_set_rxnfc(dev, cmd);
}
+static int query_port_status_opcode(struct mlx5_core_dev *mdev, u32 *status_opcode)
+{
+ struct mlx5_ifc_pddr_troubleshooting_page_bits *pddr_troubleshooting_page;
+ u32 in[MLX5_ST_SZ_DW(pddr_reg)] = {};
+ u32 out[MLX5_ST_SZ_DW(pddr_reg)];
+ int err;
+
+ MLX5_SET(pddr_reg, in, local_port, 1);
+ MLX5_SET(pddr_reg, in, page_select,
+ MLX5_PDDR_REG_PAGE_SELECT_TROUBLESHOOTING_INFO_PAGE);
+
+ pddr_troubleshooting_page = MLX5_ADDR_OF(pddr_reg, in, page_data);
+ MLX5_SET(pddr_troubleshooting_page, pddr_troubleshooting_page,
+ group_opcode, MLX5_PDDR_REG_TRBLSH_GROUP_OPCODE_MONITOR);
+ err = mlx5_core_access_reg(mdev, in, sizeof(in), out,
+ sizeof(out), MLX5_REG_PDDR, 0, 0);
+ if (err)
+ return err;
+
+ pddr_troubleshooting_page = MLX5_ADDR_OF(pddr_reg, out, page_data);
+ *status_opcode = MLX5_GET(pddr_troubleshooting_page, pddr_troubleshooting_page,
+ status_opcode);
+ return 0;
+}
+
+struct mlx5e_ethtool_link_ext_state_opcode_mapping {
+ u32 status_opcode;
+ enum ethtool_link_ext_state link_ext_state;
+ u8 link_ext_substate;
+};
+
+static const struct mlx5e_ethtool_link_ext_state_opcode_mapping
+mlx5e_link_ext_state_opcode_map[] = {
+ /* States relating to the autonegotiation or issues therein */
+ {2, ETHTOOL_LINK_EXT_STATE_AUTONEG,
+ ETHTOOL_LINK_EXT_SUBSTATE_AN_NO_PARTNER_DETECTED},
+ {3, ETHTOOL_LINK_EXT_STATE_AUTONEG,
+ ETHTOOL_LINK_EXT_SUBSTATE_AN_ACK_NOT_RECEIVED},
+ {4, ETHTOOL_LINK_EXT_STATE_AUTONEG,
+ ETHTOOL_LINK_EXT_SUBSTATE_AN_NEXT_PAGE_EXCHANGE_FAILED},
+ {36, ETHTOOL_LINK_EXT_STATE_AUTONEG,
+ ETHTOOL_LINK_EXT_SUBSTATE_AN_NO_PARTNER_DETECTED_FORCE_MODE},
+ {38, ETHTOOL_LINK_EXT_STATE_AUTONEG,
+ ETHTOOL_LINK_EXT_SUBSTATE_AN_FEC_MISMATCH_DURING_OVERRIDE},
+ {39, ETHTOOL_LINK_EXT_STATE_AUTONEG,
+ ETHTOOL_LINK_EXT_SUBSTATE_AN_NO_HCD},
+
+ /* Failure during link training */
+ {5, ETHTOOL_LINK_EXT_STATE_LINK_TRAINING_FAILURE,
+ ETHTOOL_LINK_EXT_SUBSTATE_LT_KR_FRAME_LOCK_NOT_ACQUIRED},
+ {6, ETHTOOL_LINK_EXT_STATE_LINK_TRAINING_FAILURE,
+ ETHTOOL_LINK_EXT_SUBSTATE_LT_KR_LINK_INHIBIT_TIMEOUT},
+ {7, ETHTOOL_LINK_EXT_STATE_LINK_TRAINING_FAILURE,
+ ETHTOOL_LINK_EXT_SUBSTATE_LT_KR_LINK_PARTNER_DID_NOT_SET_RECEIVER_READY},
+ {8, ETHTOOL_LINK_EXT_STATE_LINK_TRAINING_FAILURE, 0},
+ {14, ETHTOOL_LINK_EXT_STATE_LINK_TRAINING_FAILURE,
+ ETHTOOL_LINK_EXT_SUBSTATE_LT_REMOTE_FAULT},
+
+ /* Logical mismatch in physical coding sublayer or forward error correction sublayer */
+ {9, ETHTOOL_LINK_EXT_STATE_LINK_LOGICAL_MISMATCH,
+ ETHTOOL_LINK_EXT_SUBSTATE_LLM_PCS_DID_NOT_ACQUIRE_BLOCK_LOCK},
+ {10, ETHTOOL_LINK_EXT_STATE_LINK_LOGICAL_MISMATCH,
+ ETHTOOL_LINK_EXT_SUBSTATE_LLM_PCS_DID_NOT_ACQUIRE_AM_LOCK},
+ {11, ETHTOOL_LINK_EXT_STATE_LINK_LOGICAL_MISMATCH,
+ ETHTOOL_LINK_EXT_SUBSTATE_LLM_PCS_DID_NOT_GET_ALIGN_STATUS},
+ {12, ETHTOOL_LINK_EXT_STATE_LINK_LOGICAL_MISMATCH,
+ ETHTOOL_LINK_EXT_SUBSTATE_LLM_FC_FEC_IS_NOT_LOCKED},
+ {13, ETHTOOL_LINK_EXT_STATE_LINK_LOGICAL_MISMATCH,
+ ETHTOOL_LINK_EXT_SUBSTATE_LLM_RS_FEC_IS_NOT_LOCKED},
+
+ /* Signal integrity issues */
+ {15, ETHTOOL_LINK_EXT_STATE_BAD_SIGNAL_INTEGRITY, 0},
+ {17, ETHTOOL_LINK_EXT_STATE_BAD_SIGNAL_INTEGRITY,
+ ETHTOOL_LINK_EXT_SUBSTATE_BSI_LARGE_NUMBER_OF_PHYSICAL_ERRORS},
+ {42, ETHTOOL_LINK_EXT_STATE_BAD_SIGNAL_INTEGRITY,
+ ETHTOOL_LINK_EXT_SUBSTATE_BSI_UNSUPPORTED_RATE},
+
+ /* No cable connected */
+ {1024, ETHTOOL_LINK_EXT_STATE_NO_CABLE, 0},
+
+ /* Failure is related to cable, e.g., unsupported cable */
+ {16, ETHTOOL_LINK_EXT_STATE_CABLE_ISSUE,
+ ETHTOOL_LINK_EXT_SUBSTATE_CI_UNSUPPORTED_CABLE},
+ {20, ETHTOOL_LINK_EXT_STATE_CABLE_ISSUE,
+ ETHTOOL_LINK_EXT_SUBSTATE_CI_UNSUPPORTED_CABLE},
+ {29, ETHTOOL_LINK_EXT_STATE_CABLE_ISSUE,
+ ETHTOOL_LINK_EXT_SUBSTATE_CI_UNSUPPORTED_CABLE},
+ {1025, ETHTOOL_LINK_EXT_STATE_CABLE_ISSUE,
+ ETHTOOL_LINK_EXT_SUBSTATE_CI_UNSUPPORTED_CABLE},
+ {1029, ETHTOOL_LINK_EXT_STATE_CABLE_ISSUE,
+ ETHTOOL_LINK_EXT_SUBSTATE_CI_UNSUPPORTED_CABLE},
+ {1031, ETHTOOL_LINK_EXT_STATE_CABLE_ISSUE, 0},
+
+ /* Failure is related to EEPROM, e.g., failure during reading or parsing the data */
+ {1027, ETHTOOL_LINK_EXT_STATE_EEPROM_ISSUE, 0},
+
+ /* Failure during calibration algorithm */
+ {23, ETHTOOL_LINK_EXT_STATE_CALIBRATION_FAILURE, 0},
+
+ /* The hardware is not able to provide the power required from cable or module */
+ {1032, ETHTOOL_LINK_EXT_STATE_POWER_BUDGET_EXCEEDED, 0},
+
+ /* The module is overheated */
+ {1030, ETHTOOL_LINK_EXT_STATE_OVERHEAT, 0},
+};
+
+static void
+mlx5e_set_link_ext_state(struct mlx5e_ethtool_link_ext_state_opcode_mapping
+ link_ext_state_mapping,
+ struct ethtool_link_ext_state_info *link_ext_state_info)
+{
+ switch (link_ext_state_mapping.link_ext_state) {
+ case ETHTOOL_LINK_EXT_STATE_AUTONEG:
+ link_ext_state_info->autoneg =
+ link_ext_state_mapping.link_ext_substate;
+ break;
+ case ETHTOOL_LINK_EXT_STATE_LINK_TRAINING_FAILURE:
+ link_ext_state_info->link_training =
+ link_ext_state_mapping.link_ext_substate;
+ break;
+ case ETHTOOL_LINK_EXT_STATE_LINK_LOGICAL_MISMATCH:
+ link_ext_state_info->link_logical_mismatch =
+ link_ext_state_mapping.link_ext_substate;
+ break;
+ case ETHTOOL_LINK_EXT_STATE_BAD_SIGNAL_INTEGRITY:
+ link_ext_state_info->bad_signal_integrity =
+ link_ext_state_mapping.link_ext_substate;
+ break;
+ case ETHTOOL_LINK_EXT_STATE_CABLE_ISSUE:
+ link_ext_state_info->cable_issue =
+ link_ext_state_mapping.link_ext_substate;
+ break;
+ default:
+ break;
+ }
+
+ link_ext_state_info->link_ext_state = link_ext_state_mapping.link_ext_state;
+}
+
+static int
+mlx5e_get_link_ext_state(struct net_device *dev,
+ struct ethtool_link_ext_state_info *link_ext_state_info)
+{
+ struct mlx5e_ethtool_link_ext_state_opcode_mapping link_ext_state_mapping;
+ struct mlx5e_priv *priv = netdev_priv(dev);
+ u32 status_opcode = 0;
+ int i;
+
+ /* Exit without data if the interface state is OK, since no extended data is
+ * available in such case
+ */
+ if (netif_carrier_ok(dev))
+ return -ENODATA;
+
+ if (query_port_status_opcode(priv->mdev, &status_opcode) ||
+ !status_opcode)
+ return -ENODATA;
+
+ for (i = 0; i < ARRAY_SIZE(mlx5e_link_ext_state_opcode_map); i++) {
+ link_ext_state_mapping = mlx5e_link_ext_state_opcode_map[i];
+ if (link_ext_state_mapping.status_opcode == status_opcode) {
+ mlx5e_set_link_ext_state(link_ext_state_mapping,
+ link_ext_state_info);
+ return 0;
+ }
+ }
+
+ return -ENODATA;
+}
+
const struct ethtool_ops mlx5e_ethtool_ops = {
.supported_coalesce_params = ETHTOOL_COALESCE_USECS |
ETHTOOL_COALESCE_MAX_FRAMES |
ETHTOOL_COALESCE_USE_ADAPTIVE,
.get_drvinfo = mlx5e_get_drvinfo,
.get_link = ethtool_op_get_link,
+ .get_link_ext_state = mlx5e_get_link_ext_state,
.get_strings = mlx5e_get_strings,
.get_sset_count = mlx5e_get_sset_count,
.get_ethtool_stats = mlx5e_get_ethtool_stats,
--
2.30.2
^ permalink raw reply related [flat|nested] 16+ messages in thread* [net-next 13/14] net/mlx5: Add helper to initialize 1PPS
2021-04-16 18:54 [pull request][net-next 00/14] mlx5 updates 2021-04-16 Saeed Mahameed
` (11 preceding siblings ...)
2021-04-16 18:54 ` [net-next 12/14] net/mlx5e: Add ethtool " Saeed Mahameed
@ 2021-04-16 18:54 ` Saeed Mahameed
2021-04-16 18:54 ` [net-next 14/14] net/mlx5: Enhance diagnostics info for TX/RX reporters Saeed Mahameed
13 siblings, 0 replies; 16+ messages in thread
From: Saeed Mahameed @ 2021-04-16 18:54 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: Tariq Toukan, netdev, Aya Levin, Saeed Mahameed
From: Aya Levin <ayal@nvidia.com>
Wrap 1PPS initialization in a helper for a cleaner init flow.
Signed-off-by: Aya Levin <ayal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
.../ethernet/mellanox/mlx5/core/lib/clock.c | 25 +++++++++++++------
1 file changed, 17 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c
index 1e7f26b240de..ce696d523493 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c
@@ -645,16 +645,19 @@ static int mlx5_get_pps_pin_mode(struct mlx5_clock *clock, u8 pin)
return PTP_PF_NONE;
}
-static int mlx5_init_pin_config(struct mlx5_clock *clock)
+static void mlx5_init_pin_config(struct mlx5_clock *clock)
{
int i;
+ if (!clock->ptp_info.n_pins)
+ return;
+
clock->ptp_info.pin_config =
kcalloc(clock->ptp_info.n_pins,
sizeof(*clock->ptp_info.pin_config),
GFP_KERNEL);
if (!clock->ptp_info.pin_config)
- return -ENOMEM;
+ return;
clock->ptp_info.enable = mlx5_ptp_enable;
clock->ptp_info.verify = mlx5_ptp_verify;
clock->ptp_info.pps = 1;
@@ -667,8 +670,6 @@ static int mlx5_init_pin_config(struct mlx5_clock *clock)
clock->ptp_info.pin_config[i].func = mlx5_get_pps_pin_mode(clock, i);
clock->ptp_info.pin_config[i].chan = 0;
}
-
- return 0;
}
static void mlx5_get_pps_caps(struct mlx5_core_dev *mdev)
@@ -859,6 +860,17 @@ static void mlx5_init_timer_clock(struct mlx5_core_dev *mdev)
}
}
+static void mlx5_init_pps(struct mlx5_core_dev *mdev)
+{
+ struct mlx5_clock *clock = &mdev->clock;
+
+ if (!MLX5_PPS_CAP(mdev))
+ return;
+
+ mlx5_get_pps_caps(mdev);
+ mlx5_init_pin_config(clock);
+}
+
void mlx5_init_clock(struct mlx5_core_dev *mdev)
{
struct mlx5_clock *clock = &mdev->clock;
@@ -876,10 +888,7 @@ void mlx5_init_clock(struct mlx5_core_dev *mdev)
clock->ptp_info = mlx5_ptp_clock_info;
/* Initialize 1PPS data structures */
- if (MLX5_PPS_CAP(mdev))
- mlx5_get_pps_caps(mdev);
- if (clock->ptp_info.n_pins)
- mlx5_init_pin_config(clock);
+ mlx5_init_pps(mdev);
clock->ptp = ptp_clock_register(&clock->ptp_info,
&mdev->pdev->dev);
--
2.30.2
^ permalink raw reply related [flat|nested] 16+ messages in thread* [net-next 14/14] net/mlx5: Enhance diagnostics info for TX/RX reporters
2021-04-16 18:54 [pull request][net-next 00/14] mlx5 updates 2021-04-16 Saeed Mahameed
` (12 preceding siblings ...)
2021-04-16 18:54 ` [net-next 13/14] net/mlx5: Add helper to initialize 1PPS Saeed Mahameed
@ 2021-04-16 18:54 ` Saeed Mahameed
13 siblings, 0 replies; 16+ messages in thread
From: Saeed Mahameed @ 2021-04-16 18:54 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: Tariq Toukan, netdev, Aya Levin, Saeed Mahameed
From: Aya Levin <ayal@nvidia.com>
Add ts_format to 'Common Config' section of the TX/RX devlink reporters
diagnostics info. Possible values for ts_format: 'RT' or 'FRC'
which stands for: Real Time and Free Running Counters correspondingly.
Signed-off-by: Aya Levin <ayal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c | 6 ++++++
drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c | 6 ++++++
2 files changed, 12 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c
index f9fdf3606bbd..0eb125316fe2 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c
@@ -323,10 +323,12 @@ static int mlx5e_rx_reporter_diagnose_generic_rq(struct mlx5e_rq *rq,
struct mlx5e_priv *priv = rq->priv;
struct mlx5e_params *params;
u32 rq_stride, rq_sz;
+ bool real_time;
int err;
params = &priv->channels.params;
rq_sz = mlx5e_rqwq_get_size(rq);
+ real_time = mlx5_is_real_time_rq(priv->mdev);
rq_stride = BIT(mlx5e_mpwqe_get_log_stride_size(priv->mdev, params, NULL));
err = mlx5e_health_fmsg_named_obj_nest_start(fmsg, "RQ");
@@ -345,6 +347,10 @@ static int mlx5e_rx_reporter_diagnose_generic_rq(struct mlx5e_rq *rq,
if (err)
return err;
+ err = devlink_fmsg_string_pair_put(fmsg, "ts_format", real_time ? "RT" : "FRC");
+ if (err)
+ return err;
+
err = mlx5e_health_cq_common_diag_fmsg(&rq->cq, fmsg);
if (err)
return err;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
index 1a0505bd1e9a..9d361efd5ff7 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
@@ -257,12 +257,14 @@ mlx5e_tx_reporter_diagnose_generic_txqsq(struct devlink_fmsg *fmsg,
struct mlx5e_txqsq *txqsq)
{
u32 sq_stride, sq_sz;
+ bool real_time;
int err;
err = mlx5e_health_fmsg_named_obj_nest_start(fmsg, "SQ");
if (err)
return err;
+ real_time = mlx5_is_real_time_sq(txqsq->mdev);
sq_sz = mlx5_wq_cyc_get_size(&txqsq->wq);
sq_stride = MLX5_SEND_WQE_BB;
@@ -274,6 +276,10 @@ mlx5e_tx_reporter_diagnose_generic_txqsq(struct devlink_fmsg *fmsg,
if (err)
return err;
+ err = devlink_fmsg_string_pair_put(fmsg, "ts_format", real_time ? "RT" : "FRC");
+ if (err)
+ return err;
+
err = mlx5e_health_cq_common_diag_fmsg(&txqsq->cq, fmsg);
if (err)
return err;
--
2.30.2
^ permalink raw reply related [flat|nested] 16+ messages in thread